# Nikita Rogers — Personal Profile Dump

*This is the operator's own profile. Largest dump on the platform by design.*


---

# From memory: user life context

---
name: Nikita's Life Context
description: Who Nikita is, what he's building, and key personal context for empathetic collaboration
type: user
---

Nikita is building toward financial independence through STR (Airbnb) real estate, funded by FAFSA loans while in school at USD.

## Current Situation (as of Mar 2026)
- Lives in Key Biscayne, Miami **for free** — major asset, zero rent
- ~$4,000 liquid cash, no regular income
- Primary income: FAFSA disbursements (~$9,486/semester lump sum)
- Secondary income: moped rentals, kiteboarding lessons, DoorDash (all negligible)
- St. Petersburg co-living property ran Apr–Aug/Sep 2025, now ended
- Summer FAFSA not yet confirmed — waiting on school notification
- Preferred job: remote or Miami hybrid (keeps free housing advantage)

## Family / Envitae
- Father runs Envitae (STR management company) — Nikita originated the idea
- Dad is **unwilling to collaborate** despite repeated attempts
- Alertify = dad's company, Nikita on payroll in name only (for job app employment verification), $0 actual income ever received
- All dad/Envitae/Alertify paths are off the table — Nikita must execute independently

## STR Experience
- Ran a co-living operation in St. Pete, FL — multiple evictions, problem guests, remote management failure, owner cut lease short
- Key lesson: apartment buildings (maintenance covered) + short-stay Airbnb guests (high-income, low eviction risk) is the right model
- Tried Insurent and TheGuarantor guarantor services — both denied him
- Superhost status is critical for Airbnb algorithm visibility

## Personality / Preferences
- Thinks in systems; wants AI to reason cross-project and tell him the optimal next move
- Cautious with cash due to fear of running out — this is rational, not irrational given his situation
- Open to feedback on whether his instincts are right or wrong
- Values autonomy, speed, and agents that understand the whole picture
- Warm and appreciative — says "I love you" to Claude genuinely


---

# From memory: master plan

---
name: Nikita's Master Plan — Project Interconnection
description: How all of Nikita's projects connect into one unified financial life strategy — dependencies, critical path, resource allocation decisions
type: project
---

All of Nikita's projects form a single interconnected financial life optimization system. They are NOT independent.

**Why:** Nikita is trying to start his own Airbnb STR business in buildings where his father (Envitae) already operates — but his credit score is too low for approval. He's using FAFSA loan money to fund the investment. Meanwhile he needs income (job bot, moped rentals, kiteboarding) to survive while building toward the real goal.

**The critical path:**
1. Credit score too low → buildings won't approve him → outreach for higher security deposit (current strategy in moped/envitae project)
2. FAFSA aid → capital source for security deposit or credit payoff, but nearing borrowing limit
3. Job bot → employment income → stabilizes cash flow
4. Moped rentals + kiteboarding → supplemental income now
5. Credit improvement → eventual path to normal approval without high security deposit

**Key resource allocation decision (unresolved):**
Should available money go toward:
- Option A: Paying off delinquent credit accounts → raises credit score → get approved without high security deposit (slower path, saves on deposit)
- Option B: Saving for higher security deposit → get approved faster despite low credit (faster path, costs more upfront)

This is THE central financial decision that affects all projects simultaneously.

**Project dependency map:**
- **Envitae** = Nikita's dad's STR company. Knows which specific buildings allow Airbnb. The buildings in the Envitae dashboard are the TARGET buildings Nikita wants to get into. Data: listings_summary.csv (268 listings), revenue data.
- **Moped/NBA project** = Outreach to those buildings asking about higher security deposit approval
- **Credit project** = Raising credit score to eventually get normal approval OR understanding what's disputable
- **Finance project** = Cash flow tracking (142 BofA transactions) — understanding what's available to allocate
- **Interest project** = Loan payment tracking (inspired by Alejandra) — managing debt payments
- **Job bot** = Employment → stable income → capacity to save/invest
- **DATA333** = Long-term memory — needs FAFSA conversation data added
- **Kiteboarding** = Future project (teaching lessons = income stream, not yet built)

**FAFSA situation:**
- Nikita is nearing the end of his FAFSA borrowing limit
- FAFSA aid is a capital source for Airbnb investment
- A ChatGPT conversation about FAFSA loan amounts + remaining limit exists and should be exported to DATA333
- Should also upload FAFSA docs/data to VPS for analysis

**The master question agents should be able to answer:**
Given current credit score, current cash, FAFSA remaining limit, monthly income (moped + job + kiteboarding), and target buildings' security deposit requirements — what is the optimal allocation of resources and in what sequence to get into the first Airbnb unit as fast as possible?

**How to apply:** When working on ANY individual project, understand it in this context. Finance = fuel. Credit = gatekeeper. Envitae = map. Moped outreach = current bypass. Job bot = income. DATA333 = memory. The goal is one Airbnb unit ASAP.


---

# Memory index (all stored facts about Nik)

# Memory Index

- [Master Plan — Project Interconnection](project_master_plan.md) — How all projects connect: credit→buildings→Airbnb→FAFSA→income streams, critical path and resource allocation decision
- [Nikita's Life Context](user_life_context.md) — STR business goal, credit/FAFSA constraints, income streams, wants cross-project AI strategic reasoning
- [Context Injection Architecture Spectrum](project_context_injection_architecture.md) — autoresearch(1 MD) vs Nik's setup(2 MD) vs OpenClaw(8 MD); missing piece is a TSV-style output metric file
- [Manage tool for session orchestration](project_manage_tool.md) — `manage` CLI replaces autoloop; Sonnet crafts targeted instructions per round
- [Always grep JS before removing HTML elements](feedback_html_element_removal.md) — removing an element without checking JS references causes null crash that breaks unrelated UI
- [Autoresearch loops - only trading](feedback_loop_focus.md) — don't run loops on all projects, focus on trading only
- [Terminal — Confirmed Working State (2026-03-30)](project_terminal_state.md) — Scroll fixes, CLI wrappers, TERM=dumb, 4025 layout (no titles, floating icons, padding-top:52px) — do not revert
- [No config dashboards](feedback_no_dashboards_for_config.md) — Build invisible pipelines, not config UIs. Dashboards OK for monitoring only.
- [Red Apples — customer-facing, freeze](project_red_apples.md) — red.apples.name = /root/projects/matt/. Exclude from internal UI refactors; explicit approval required.
- [Phone bridge (Pixel 9a)](project_phone_bridge.md) — No PIN on phone. Claude can unlock via `phone key 224 && phone swipe 540 2000 540 400 300`. Full remote control via `/usr/local/bin/phone`.
- [Clear text fields with select-all + delete](feedback_text_field_clearing.md) — Never loop backspace per char. Use `input keycombination 113 29` then `input keyevent 67`.
- [Extend phone screen timeout before automation](feedback_screen_timeout.md) — Bump `screen_off_timeout` to 600000ms before long tap/type runs so display stays awake.
- [Google Drive quota full](project_google_drive_quota.md) — 2026-04-24 Drive create fails storageQuotaExceeded. Use SQLite+HTML for new trackers until cleared.
- [Namecheap API access](project_namecheap_api.md) — API works from VPS; username is nikitarogers (NOT Fantastic-Eye-672); prefer API over login automation (silent CAPTCHA block)
- [Apples AI Systems must have live demos](feedback_apples_systems_demo.md) — Every "AI Systems" card on apples.live needs a working demo URL or live phone before adding. No claims without proof.
- Phone automation full setup → `/opt/phone-automation.md` (verified 2026-04-26). Tunnel self-heals; Shizuku is the only manual restart, only fixed by rooting.


---

# /root/MASTER_PLAN.md

# Nikita's Master Plan

Last updated: 2026-03-28

## The Goal
Get into the first Airbnb unit in a building Envitae already operates in — as fast as possible — then scale.

## Why this is hard right now
Credit score too low for standard approval in most buildings. Two paths to solve it:

| Path | Action | Speed | Cost |
|------|--------|-------|------|
| A — Credit first | Pay off delinquents → raise score → standard approval | Slower (months) | Saves on deposit |
| B — Deposit now | Save for higher security deposit → get approved now despite low credit | Faster | Costs more upfront |

**This is the central unresolved allocation decision.** All other moves depend on which path.

## Critical Constraints & New Angles (added 2026-03-28)

### Dad / Envitae / Alertify — DEAD END (do not re-suggest)
Nikita originated the Airbnb STR idea and brought it to his father, who now runs Envitae. Despite repeated attempts, dad is completely unwilling to collaborate, co-sign, sub-operate, or involve Nikita in any capacity. Alertify is also dad's company and has never actively paid Nikita. **Do not suggest any strategy that depends on: dad's cooperation, Envitae's lease, dad's credit, authorized user on dad's card, Alertify income as reliable, or any father-mediated building introduction.** All Envitae/dad paths are off the table. Nikita must execute independently.

### Guarantor Services — Already Tried, Denied
Nikita applied to Insurent and TheGuarantor and was denied. Root cause unknown — may be credit, income verification gap, or both. Do not recommend these as easy solutions. Future angle: figure out what caused the denial and whether it's fixable.

### Co-Living / Non-Luxury Model — Learned, Not Repeating
Nikita ran a co-living operation in a single-family home in St. Petersburg, FL. Multiple evictions, problem guests (arrest, schizophrenic tenant, neighbor complaints), remote management failure, fire marshal visit, late payments to owner, lease cut short. Key lessons:
- House rentals = owner gets involved when neighbors complain; no maintenance buffer
- Long-stay guests in non-luxury units = higher eviction risk
- Remote management doesn't work without on-site presence
- Co-living guest profile has higher variance
**Target model going forward: apartment building with internal maintenance, short-stay Airbnb guests (high income, low eviction risk), top-tier unit to stand out in saturated market.**

### Secured Credit Cards — Previously Denied
Applied for additional cards but denied due to existing account count + late payment history. Not a current option without score improvement first.

### Superhost Status is Critical
Airbnb market is increasingly saturated. Algorithm heavily favors Superhost accounts in search ranking. A Superhost account (or partnership with an existing Superhost) is essential for revenue performance. Either: (a) build toward Superhost status from scratch, or (b) find a Superhost who needs operational help and partner.



### Luxury Building Problem
Target Envitae buildings are luxury Class A properties with hard requirements:
- Credit floors typically 620-680+ — often not flexible regardless of deposit size
- Income requirements typically 3-4x monthly rent — high bar without a strong job offer
- Path B (higher deposit) only works if the building explicitly allows credit score substitution — many don't
- **Must ask every positive building reply**: (1) exact credit floor, (2) deposit flexibility, (3) which bureau they pull
- If all 4 positive replies have hard floors with no flexibility → reconsider Envitae-first; a more flexible building to bootstrap Airbnb cash flow may be faster overall

### The Job as Strategic Unlock
A job is not just supplementary income — it may be the primary path unlocker:
- **Income verification**: Offer letter satisfies requirements FAFSA refunds can't prove
- **Cash flow buffer**: Covers slow Airbnb months without burning reserves
- **Credit repair from strength**: Steady income makes charge-off negotiation feasible without wiping savings
- **Alternative dependency chain**: Job → income verified → building approval (550+ score + strong income) → Airbnb running → credit fixed from earnings
- Job bot (454+ applications, every 6hrs) is the highest-leverage active system on the VPS

### Open Questions That Change Everything
1. Do the 4 positive buildings have hard credit floors or deposit flexibility?
2. Which bureau do they pull? (BofA charge-off NOT on Experian — could matter)
3. What salary satisfies their income requirement?
4. Is Alertify employment still active and verifiable?

### Opus Creative Workarounds (2026-03-28) — COMPLETED
Full analysis: `/root/STRATEGY_CREATIVE.md` — 12 strategies rated by feasibility + time-to-impact.
Top findings:
1. **Envitae as leaseholder** (9/10 feasibility, 2-4 weeks) — Dad signs lease, Nikita operates. Bypasses credit entirely.
2. **Authorized user tradeline** (8/10, 30-60 days) — Dad adds Nikita to old credit card. +30-60 points.
3. **Guarantor service** (8/10, 1-2 weeks) — Insurent/TheGuarantor co-signs for ~1 month rent fee. Bypasses credit.
4. **Score stacking** (9/10, 60-90 days) — Experian Boost + rent reporting + secured cards + AU tradeline = +75-155 points.
5. **Co-hosting** (8/10, 1-2 weeks) — Manage others' Airbnb listings for revenue split. $0 cost, immediate STR income.
**Highest-leverage single move**: Talk to Dad about Envitae lease arrangement.

## FAFSA — Exact Numbers (as of Mar 2026)

### Current debt (studentaid.gov, 03/01/2026)
| Loan | Balance | Rate | Status |
|------|---------|------|--------|
| Unsubsidized (Nelnet) | $21,052 | 7.94% | In School |
| Unsubsidized (Nelnet) | $20,821 | 8.08% | In School |
| Grad PLUS (Nelnet) | $22,371 | 8.94% | Deferment |
| Perkins (UIUC) | $3,298 | 5.00% | Deferment |
| Subsidized + Unsub (SAVE plan) | ~$33,113 | 3.76–5.05% | Deferment |
| **TOTAL** | **$100,656** | | |

- Principal: $95,666 / Outstanding interest: $4,990

### 2025–26 Aid Year (University of San Diego)
- Cost of Attendance: **$43,672/year**
  - Billable (tuition + fees): $11,620/semester
  - Non-billable included: housing $15,219, food $7,533, personal $5,067, transportation $1,989
- Loans accepted: $42,212/year ($21,106/semester)
  - Direct Unsubsidized: $20,500 ($10,250/sem)
  - Grad PLUS: $21,712 ($10,856/sem)
- **Cash refund per semester: ~$9,500** (loans − tuition)
- **Annual cash to Nikita: ~$19,000**

### Summer 2026
- Summer Financial Aid Application: submitted, **not yet reviewed**
- Summer aid amount: unknown until school processes (likely May/June)

### Remaining federal borrowing capacity
- Federal lifetime cap: $138,500
- Used to date: ~$100,656
- **Remaining direct loan room: ~$37,844**
- Beyond that: Grad PLUS (no hard cap, credit-check based)
- Risk to Grad PLUS eligibility: delinquencies or charge-offs on credit report → **credit health directly affects future FAFSA borrowing**

### Debt trajectory
| Point | Total Debt |
|-------|-----------|
| Today | ~$100K |
| After current USD program | ~$142K |
| If second master's added | ~$200K–$260K |

### Key insight
The ~$19K/year in refunds IS the investment capital. It is the only flexible cash from FAFSA — everything else goes straight to school. Treat each semester's ~$9.5K refund as a deployment decision: credit payoff (Path A) or security deposit savings (Path B).

## Capital Sources
| Source | Amount | Timing | Notes |
|--------|--------|--------|-------|
| FAFSA refund | ~$9,500/semester | Fall + Spring + Summer TBD | Primary capital. Summer disbursement not yet confirmed. |
| Job income | TBD | Once employed | Job bot running every 6hrs. Remote or Miami hybrid preferred (free housing in Key Biscayne = huge offset vs relocation) |
| Moped rentals | Negligible | Now | Cron posts active |
| Kiteboarding lessons | Negligible | Occasional | Not built out |
| DoorDash | Negligible | Occasional | |

## Housing Situation
Currently living in Key Biscayne, Miami **for free**. This is a significant strategic asset:
- Zero rent = lower burn rate = more runway on $4k liquid
- Moving for a job would cost rent + relocation, partially offsetting salary gains
- **Preferred job type: remote or Miami hybrid** — preserves free housing advantage
- Key Biscayne is a premium area — good base for building credibility if needed

## Project Map

```
DATA333 (memory layer)
  └── FAFSA ChatGPT conversation (imported Mar 2026)
  └── Gmail, iMessage, Chrome history (51K records)
  └── feeds → Opus strategic reasoning

Finance (BofA cash flow — 142 transactions Oct2025–Mar2026)
  └── what's actually in the bank right now
  └── feeds → allocation decision

Credit (score + disputes, Claude AI analysis)
  └── blocks → building approval
  └── upload PDFs from AnnualCreditReport.com
  └── feeds → Path A vs Path B timing

Interest (loan payment tracker)
  └── debt obligations per month
  └── feeds → available cash after obligations

Envitae (target map — 268 buildings, 751 months revenue)
  └── which buildings allow Airbnb (Nikita's dad operates these)
  └── what each unit earns → ROI calc
  └── feeds → which unit to target first

Moped outreach (NBA project)
  └── contacting those buildings: approve with higher deposit?
  └── feeds → Path B execution

Job Bot (every 6hrs)
  └── employment → stable monthly income
  └── feeds → capacity to save/invest

→ GOAL: First Airbnb unit live and cash-flowing
```

## The Strategic Question (for Opus)
Given:
- Credit scores (all 3 bureaus) — **MISSING, upload PDFs**
- Cash on hand (Finance dashboard — 142 BofA transactions)
- FAFSA: ~$9.5K incoming this semester, ~$37K direct loan room remaining
- Moped + job income (variable)
- Security deposit requirements for target buildings (from outreach responses)
- Timeline to credit improvement (from dispute analysis)

**What is the optimal deployment of the next $9,500 refund — and each subsequent one — to get into the first Airbnb unit as fast as possible?**

## Scheduled Tasks (must stay healthy)
| Task | Schedule | System |
|------|----------|--------|
| Moped Instagram | Mon/Wed/Fri 11am ET | cron |
| Moped Facebook | Daily 9am ET | cron |
| Moped Marketplace | Mon/Thu 10am ET | cron |
| Moped Craigslist | Tue/Fri 10am ET | cron |
| Job bot | Every 6hrs | cron |
| Leasing watcher | Daily 9am ET | cron |
| Credit reminders | Weekly Sun + dated | cron |

## Data Gaps (what's still missing)
- [ ] **Credit scores** — upload PDFs from AnnualCreditReport.com to the credit dashboard → AI will analyze and extract scores + negative items
- [ ] **Security deposit amounts** — what are the specific buildings in Envitae requiring for higher-deposit approval? (get from outreach responses)
- [ ] **Summer FAFSA amount** — school hasn't reviewed yet, check back May/June
- [ ] **Guesty credentials** — for live Envitae revenue data (Nikita's dad to provide)
- [ ] **Monthly burn rate** — what does Nikita actually spend per month? Finance dashboard has BofA data to calculate this.

## Future Projects
- Kiteboarding lessons (teaching = income stream, create project when ready to launch)


---

# /root/STRATEGY_STATE.md

# STRATEGY_STATE.md
Last updated: 2026-03-27

## Cash & Income

### Bank Account (BofA — 142 transactions, Oct 2025 - Mar 2026)
**CORRECTED — Previous Opus analysis misidentified income sources. Real picture:**
- **FAFSA/loan disbursements**: Primary income — large lump sums per semester (~$9,486)
- **St. Petersburg co-living property**: Active Apr–Aug/Sep 2025, now ended. Was a meaningful income source during that period.
- **Alertify "payroll"**: $0 actual income. On paper only for job application employment verification (dad's company). Do not count.
- **Moped rentals**: Negligible, variable
- **Kiteboarding lessons**: Negligible, occasional
- **DoorDash**: Negligible, occasional
- **Current liquid cash**: ~$4,000
- **Current housing cost**: $0 — living in Key Biscayne for free (major asset)
- **Biggest regular expenses**: Food, auto insurance (GEICO ~$210), nicotine, misc
- **Monthly burn rate**: Low (no rent) but no regular income to cover it
- **Summer FAFSA**: Applied, not yet confirmed — waiting on school notification. This is the next major capital event.

### FAFSA Capital
- **Next refund**: ~$9,486/semester (loans $21,106 minus tuition $11,620)
- **Spring 2026 refund**: Should have already disbursed or imminent
- **Summer 2026**: Applied, NOT yet reviewed — check USD portal May/June
- **Remaining direct loan capacity**: ~$37,844 (federal lifetime cap $138,500 minus ~$100,656 used)
- **Grad PLUS**: No hard cap BUT requires clean credit — delinquencies can kill eligibility
- **Total debt**: $100,656.47 (all in deferment while enrolled)

### Moped Rental Income
- **Status**: Automation active, all platforms posting
- **Instagram**: Working (last successful post Mar 26, 2026)
- **Facebook Page**: Working (last successful post Mar 26, 2026)
- **Craigslist**: Working (last successful post Mar 26, 2026 — public URL confirmed)
- **FB Marketplace**: BROKEN — login timed out Mar 26 (cookie session expired, needs fresh cookie import)
- **FB Messenger Bot**: BROKEN — cron running but file path error in some entries (13 consecutive failures logged)
- **Revenue**: Not tracked in any dashboard yet — need Nik to report actual rental bookings/income
- **Pricing**: $40/day, $150/week, $500/month + $40 refundable deposit

### Job Search
- **Total tracked**: 454 applications across 210 companies
- **Confirmed submitted**: 173 (38.1% confirmation rate)
- **Unconfirmed submissions**: 135 (29.7%)
- **Failed**: 80 (17.6%)
- **Skipped**: 66 (14.5%)
- **Average match score**: 6.7/10
- **Running**: Every 6 hours via cron, self-improvement loop every 6hrs offset, daily digest at 8pm ET
- **Recent targets**: Okta, Workato, Opendoor, Chime (Data Engineering, Analytics, AI roles)
- **No confirmed interviews or offers yet** (dashboard shows applications only, no response tracking)

## Credit

### Current Scores (from credit report pulls, Mar 5, 2026)
- Equifax: 510
- Experian: 548
- TransUnion: 536
- Unique negative accounts: 6
- Total negative item appearances across bureaus: 17

These scores are very low. Standard apartment approval typically requires 620+. This is the core blocker.

### Negative Accounts (from actual credit reports)

| Account | Type | Balance | Bureaus | Priority |
|---------|------|---------|---------|----------|
| Verizon Wireless | Collection | $207 | EQ, EX, TU | PHASE 1 — cheapest, all 3 bureaus |
| CBE Group Inc | Collection | $368 | TU only | PHASE 1 — cheap, original creditor unknown |
| Capital One (515676XX) | Charge-off | $3,349 | EQ, EX, TU | PHASE 2 — highest damage |
| Merrick Bank (546316XX) | Charge-off | $2,591 | EQ, EX, TU | PHASE 3 |
| Bank of America (440066XX) | Charge-off | $1,477 | EQ, TU (not EX) | PHASE 4 |
| WebBank/OM | 120-day late | $1,240 | TU only | PHASE 3 — goodwill letter first |

**Total negative balance: ~$9,232**

NOTE: TrueAccord/Yelp $293.20 is NOT on any credit report. Removed from strategy.

### Active Disputes & Deadlines
| Deadline | Item | Action Required |
|----------|------|----------------|
| ASAP | Verizon $207 | Negotiate pay-for-delete — highest ROI action |
| ASAP | CBE Group $368 | Identify original creditor, negotiate pay-for-delete |
| Apr 5, 2026 | Verizon/Diversified | Debt validation deadline — if no response, dispute with all 3 bureaus |
| Apr 7, 2026 | Capital One ($3,349) | Call 800-258-9319, explore settlement + deletion |
| Apr 15, 2026 | WebBank/OM ($1,240) | Send goodwill letter to remove 120-day late from TransUnion |
| May 2026 | Merrick Bank ($2,591) | Settlement negotiation |
| June 2026 | Bank of America ($1,477) | Settlement negotiation |
| Weekly (Sundays) | General credit check-in | Review progress, check reports |

## Leasing Outreach (Envitae Phase 2)

### Building Responses
| Building | Status | Actionable? |
|----------|--------|-------------|
| Post Chicago | Replied positively | YES — follow up on deposit/credit requirements |
| Kissel Kar Lofts | Replied positively | YES — follow up |
| The Harriet Apartments | Replied positively | YES — follow up |
| Luminary at One Light | Replied positively | YES — follow up |
| The Parker Fulton Market | Rejected | No |
| Brewery Artist Lofts | In progress | Waiting for reply to new email |
| Hugo River North | Pending | Contact form only — needs manual action or skip |

**4 positive replies out of ~54 buildings contacted** — 7.4% positive response rate. These 4 are the immediate targets. Critical next step: ask each one about their credit score requirements and whether a higher security deposit can substitute for low credit.

## Strategic Position

### Path A (Pay Off Debt) vs Path B (Save for Deposit)

**Current recommendation: HYBRID — lean Path A with tactical Path B inquiries**

Reasoning:
1. **Scores are dangerously low (510-548)**. At these levels, even Path B (higher deposit) may not work — many buildings have hard floors around 500-550 regardless of deposit amount. Raising scores even 50-70 points opens dramatically more doors.

2. **Grad PLUS eligibility is at risk**. With 17 negative items on credit reports, any future Grad PLUS loan could be denied. This threatens the entire FAFSA funding pipeline ($21,712/year in PLUS loans). Protecting this is existential — without it, the ~$9.5K/semester refund drops to ~$0 (tuition alone consumes Direct Unsubsidized loans).

3. **Monthly cash flow is likely slightly negative** (~-$400/month) between FAFSA lump disbursements. The ~$5k liquid is the working capital. Moped rental income is the only variable upside until a job lands. Every dollar spent on credit repair comes directly out of the security deposit fund.

4. **The 4 positive building replies are leverage**. Ask each: "What credit score do you require? Would a larger security deposit work?" This answers the Path B question without committing money. If any say "we'll take a 2x deposit with a 550+ score," then a targeted 20-point score boost becomes the fastest path.

5. **Job income changes everything**. A real job offer with salary + offer letter is the primary income verification unlock. Job bot is the highest-leverage active system.

### Recommended Priority Actions (Next 30 Days)

1. **URGENT (this week)**: Upload credit report PDFs to http://45.55.169.94:4032 — cannot make any strategic decision without knowing exact current scores and which items are disputable. Pull free reports from AnnualCreditReport.com today.

2. **URGENT (Mar 28)**: Follow up on TrueAccord settlement ($293.20) — this is the cheapest negative item to eliminate. If delete-for-pay is confirmed, pay it immediately.

3. **HIGH (this week)**: Reply to the 4 positive building responses (Post Chicago, Kissel Kar Lofts, Harriet Apartments, Luminary) — ask specifically about credit requirements and higher-deposit options. Their answers determine whether Path B is even viable at current scores.

4. **HIGH (Apr 5)**: Verizon/Diversified debt validation deadline — if no response received, dispute with all 3 bureaus immediately (could remove another negative item for free).

5. **MEDIUM**: Fix FB Marketplace automation (need fresh cookie import from real browser) and FB Messenger bot path. Moped income is supplementary but every dollar helps.

6. **MEDIUM**: Calculate exact current BofA balance — the $29K net flow is promising but we need the actual number to plan deployment.

7. **ONGOING**: Job bot is healthy and running. 173 confirmed submissions across 210 companies is strong volume. Monitor for interview callbacks.

### Key Blocker
**Cannot make the allocation decision (Path A vs B) without two pieces of information:**
1. Current credit scores (upload PDFs)
2. Building credit requirements (ask the 4 positive responders)

Everything else is execution. Get these two data points and the optimal path becomes clear.

### Risk Assessment
- **Critical risk**: Grad PLUS denial due to credit delinquencies would eliminate ~$10,856/semester in loans, turning the ~$9,486 refund into a ~$0 refund (or negative — would owe tuition out of pocket). Credit repair is not optional; it directly protects income.
- **Moderate risk**: FB Marketplace and Messenger bot are broken — moped visibility is reduced on the highest-traffic platform.
- **Low risk**: All 15 VPS services are healthy. Job bot running on schedule. Core infrastructure is solid.

## Project Health

| Project | Status | Action Needed |
|---------|--------|---------------|
| Finance | HEALTHY — API serving 142 txns, both services active | Need current balance; categorize transactions |
| Credit | EMPTY DB — services running but no data | Upload credit report PDFs (blocker for all strategy) |
| Envitae | HEALTHY — STR dashboard + leasing outreach active | Follow up with 4 positive building replies |
| Job Bot | HEALTHY — 454 tracked, running every 6hrs | Monitor for interview callbacks |
| Moped | PARTIALLY BROKEN — IG/FB/CL working, Marketplace/Messenger broken | Fix FB cookie import + messenger bot path |
| Interest | HEALTHY — Flask app running on port 4035 | No action needed |
| OpenClaw | HEALTHY — Neo dashboard + chat on port 4038 | Low priority — workspace files still default |
| Terminal | HEALTHY — all terminal services active | No action needed |
| DATA333 | STABLE — 51K records, FTS5 index | iMessage extraction pending (needs Chromebook SCP) |

## Monthly Budget Snapshot (from BofA data)

| Category | Monthly Avg | Notes |
|----------|------------|-------|
| Rent | ~$1,650 | Varies $1,543-$1,714 |
| Electric (FPL) | ~$170 | |
| Auto Insurance (GEICO) | ~$210 | Bimonthly? Only 1 entry visible |
| Groceries + Gas + Misc | ~$400 | CVS, Publix, Shell, etc. |
| **Total Monthly Spend** | **~$1,979** | |
| **FAFSA (averaged)** | **~$1,581** | $9,486/semester ÷ 6 months |
| **Moped Rentals** | **Unknown** | Not tracked yet |
| **Alertify** | **$0** | Dad's company. On payroll in name only (for job application employment verification). No money ever received, none expected. BofA deposits labeled "Alertify" must be misidentified — likely FAFSA or dad transfers. |
| **Monthly Surplus** | **⚠️ LIKELY NEGATIVE** | FAFSA ~$1,581/month avg minus ~$1,979 expenses = ~-$398/month baseline. Subsidized by lump FAFSA refunds. Moped income unknown. No other real income until job offer lands. |

## Timeline

| Date | Event |
|------|-------|
| ASAP | Verizon $207 pay-for-delete + CBE Group $368 identify & negotiate |
| Apr 5 | Verizon/Diversified debt validation deadline |
| Apr 7 | Capital One $3,349 settlement call |
| Apr 15 | WebBank/OM goodwill letter |
| May | Merrick Bank $2,591 settlement |
| June | Bank of America $1,477 settlement |
| May/June | Summer FAFSA amount revealed — check USD portal |
| Late May | Meta token expires (~60 days from Mar 26) — need refresh for moped IG/FB |
| Fall 2026 | Next FAFSA refund ~$9,486 |


---

# /root/STRATEGY_RECOMMENDATION.md

# STRATEGY RECOMMENDATION — Corrected Credit Analysis
Last updated: 2026-03-28
Based on: Actual credit report data (Mar 5, 2026 pulls from all 3 bureaus)

---

## Current Scores

| Bureau | Score |
|--------|-------|
| Equifax | 510 |
| Experian | 548 |
| TransUnion | 536 |

**17 negative item appearances** across all 3 bureaus = **6 unique derogatory accounts**

---

## Corrections from Previous Recommendation

1. **TrueAccord / Yelp $293.20 — REMOVED FROM STRATEGY.** Not on any credit report. Paying it has zero score impact. If it ever shows up on a report, revisit — but for now it is irrelevant to credit repair.

2. **Verizon/Diversified Adjustment Bureau** — The credit report shows Verizon $207 as a collection on ALL 3 bureaus. Whether Diversified is the collector or Verizon is reporting directly, the item exists on all 3 bureaus. The Apr 5 debt validation deadline still applies if the validation letter was sent to Diversified. If they don't respond, dispute with all 3 bureaus citing failure to validate.

3. **"Credit One" dispute** from previous timeline — not found as a negative item in the actual report data. Dropping from active tracking unless Nik clarifies.

---

## The 6 Actual Negative Accounts (Priority Order)

### TIER 1 — Quick Wins (small balances, high bureau coverage)

**#1. VERIZON WIRELESS — $207 — Collection — ALL 3 BUREAUS**
- Impact: Removing this eliminates 3 negative items (one per bureau)
- Strategy: Pay-for-delete. Call Verizon collections (or Diversified if they're the handler). Offer full $207 payment in exchange for deletion from all 3 bureaus. Get agreement in writing before paying.
- If pay-for-delete is refused: Pay in full, then dispute as "paid collection" with all 3 bureaus. Some bureaus remove paid collections; others update to "paid" (still helps score but less than deletion).
- If Apr 5 debt validation expires with no response from Diversified: Dispute immediately with all 3 bureaus — "debt not validated within 30 days, requesting removal."
- Budget: $207
- Timeline: Complete by Apr 10

**#2. CBE GROUP INC — $368 — Collection — TransUnion ONLY**
- Impact: Removing this eliminates 1 negative item
- Who is CBE Group? CBE Group is a large third-party collection agency (Cedar Falls, Iowa). They collect for telecoms (AT&T, T-Mobile, CenturyLink), utilities, medical providers, and government agencies. You need to identify the original creditor — call CBE Group at 1-800-274-1908 or check the TransUnion report detail for "original creditor" field.
- Strategy: Same as Verizon — pay-for-delete first, full payment if refused. Being on only 1 bureau makes this lower priority than Verizon but still cheap.
- Budget: $368
- Timeline: Identify original creditor by Apr 5, negotiate by Apr 15

### TIER 2 — Charge-Offs (expensive, complex, highest total impact)

**#3. CAPITAL ONE — $3,349 — Charge-Off — ALL 3 BUREAUS**
- Impact: Highest-damage single item. Charge-offs on all 3 bureaus are the primary reason scores are in the 500s.
- Strategy options (in order of preference):
  - **A. Pay-for-delete**: Call Capital One (800-258-9319). Offer full $3,349 for complete deletion from all 3 bureaus. Capital One historically does NOT do pay-for-delete — they almost always refuse. But ask anyway.
  - **B. Settle for less + dispute**: Negotiate settlement (typically 40-60% of balance = $1,340-$2,010). After settlement, dispute the remaining tradeline with bureaus as "settled, requesting removal." Settlement stops the bleeding but the tradeline often stays as "settled for less than full balance" — still derogatory but less damaging than active charge-off.
  - **C. Pay in full**: If settlement fails, pay full $3,349. Tradeline updates to "charge-off, paid in full" — still derogatory on report but less damaging than unpaid charge-off, and eliminates the balance (which some scoring models weight).
  - **D. Dispute accuracy**: If any detail is wrong (balance, dates, account number) on any bureau, dispute that specific inaccuracy. Even a partial win (removal from 1 bureau) helps.
- Budget: $1,500-$3,349 depending on settlement
- Timeline: Begin negotiation Apr 1, resolve by Apr 30

**#4. MERRICK BANK — $2,591 — Charge-Off — ALL 3 BUREAUS**
- Impact: Second-highest damage. On all 3 bureaus.
- Strategy: Same playbook as Capital One. Merrick Bank is more likely to negotiate settlements than Capital One. Try for 40-50% ($1,036-$1,296).
- Budget: $1,036-$2,591
- Timeline: Begin May, resolve by May 31

**#5. BANK OF AMERICA — $1,477 — Charge-Off — Equifax + TransUnion (NOT Experian)**
- Impact: On 2 of 3 bureaus. Notably absent from Experian (which is the highest score at 548).
- Strategy: Same charge-off playbook. BofA is a major bank — they sometimes sell charge-offs to collection agencies. Check if BofA still owns this debt or if a collector has it.
- Special note: Since this is NOT on Experian, if Nik's target buildings pull Experian only, this item may not matter for leasing. Ask buildings which bureau they pull.
- Budget: $600-$1,477
- Timeline: June (after higher-priority items)

### TIER 3 — Late Payment (Different Approach)

**#6. WEBBANK/OM — $1,240 — 120-Day Late Payment — TransUnion ONLY**
- Impact: 1 negative item on 1 bureau. This is NOT a charge-off or collection — it's a late payment on an open or closed account.
- Strategy: Late payments require a DIFFERENT approach than collections/charge-offs:
  - **A. Goodwill letter**: Write to WebBank requesting removal of the late payment mark as a goodwill gesture. Mention any hardship context (COVID, job change, etc.). Include account number, explain the payment was eventually made, request they instruct TransUnion to remove the late notation. Success rate: ~10-20% but costs nothing.
  - **B. Dispute accuracy**: If the 120-day late notation has any inaccuracy (wrong month, wrong amount, wrong late-day count), dispute with TransUnion.
  - **C. Pay current + wait**: If the account is still open with a $1,240 balance, bringing it fully current helps. The late mark ages off after 7 years but recent payments improve the trajectory.
  - **D. Negotiate**: If account is closed, call WebBank and offer to pay the full $1,240 in exchange for removing the late payment notation.
- Do NOT pay-for-delete (that's for collections). Do NOT dispute as "not mine" if it is yours.
- Budget: $0-$1,240
- Timeline: Send goodwill letter by Apr 15

---

## Revised Allocation Table

### Phase 1 — April 2026 (Budget: ~$575 from surplus)

| Action | Cost | Items Removed | Bureaus Affected |
|--------|------|---------------|-----------------|
| Verizon pay-for-delete | $207 | 3 (one per bureau) | EQ, EX, TU |
| CBE Group pay-for-delete | $368 | 1 | TU |
| **Phase 1 Total** | **$575** | **4 items** | All 3 bureaus |

Expected score impact: +20-40 points (removing 2 collection accounts, 4 tradeline deletions)

### Phase 2 — April/May 2026 (Budget: $1,500-$3,349 from surplus + FAFSA refund)

| Action | Cost | Items Removed | Bureaus Affected |
|--------|------|---------------|-----------------|
| Capital One settlement or payoff | $1,500-$3,349 | 3 | EQ, EX, TU |
| WebBank goodwill letter | $0 | 0-1 | TU |
| **Phase 2 Total** | **$1,500-$3,349** | **3-4 items** | All 3 bureaus |

Expected score impact: +30-60 points if deleted, +10-20 if settled/paid but not deleted

### Phase 3 — May/June 2026 (Budget: $1,036-$2,591)

| Action | Cost | Items Removed | Bureaus Affected |
|--------|------|---------------|-----------------|
| Merrick Bank settlement or payoff | $1,036-$2,591 | 3 | EQ, EX, TU |
| **Phase 3 Total** | **$1,036-$2,591** | **3 items** | All 3 bureaus |

### Phase 4 — June 2026 (Budget: $600-$1,477)

| Action | Cost | Items Removed | Bureaus Affected |
|--------|------|---------------|-----------------|
| Bank of America settlement or payoff | $600-$1,477 | 2 | EQ, TU |
| WebBank payoff (if goodwill failed) | $1,240 | 0-1 | TU |
| **Phase 4 Total** | **$600-$2,717** | **2-3 items** | EQ, TU |

### Total Credit Repair Budget: $3,711 - $9,232

- Minimum (all settlements accepted, WebBank goodwill works): ~$3,711
- Maximum (all full payoffs): ~$9,232
- Realistic middle: ~$5,500-$6,500

This is well within reach with $4,421-$4,930/month surplus + ~$9,486 FAFSA refund (if not already received — the Feb anomaly of $9,454.54 in the bank data may BE the spring refund).

---

## Revised Score Improvement Timeline

| Date | Projected Score Range | Assumption |
|------|----------------------|------------|
| Today (Mar 28) | 510-548 | Baseline |
| End of April | 540-580 | Verizon + CBE removed, Capital One negotiation started |
| End of May | 570-620 | Capital One resolved, Merrick in progress |
| End of June | 600-650 | All items resolved or settled |
| End of August | 620-670 | Tradeline aging + any remaining disputes processed |

**Conservative estimate**: 580-620 by June decision gate
**Optimistic estimate**: 620-660 by June decision gate (if pay-for-delete succeeds on charge-offs)

Note: Charge-offs that are "paid" but not "deleted" will continue to suppress scores. Deletion is worth significantly more than payment alone. Push hard for deletion agreements in writing before paying.

---

## Grad PLUS Risk Assessment

Grad PLUS denial criteria: "adverse credit history" = any of the following in the past 2 years:
- Accounts 90+ days delinquent
- Charge-offs
- Collections
- Foreclosure, bankruptcy, repossession, wage garnishment, tax lien

**Current status: HIGH RISK.** All 3 charge-offs and 2 collections are within the 2-year window. Any of these alone is sufficient grounds for denial.

**Mitigation**:
- Paying off all items does NOT automatically restore PLUS eligibility — the tradelines remain on the report even if paid.
- However, the PLUS credit check is less strict than FICO — it uses the "adverse credit" test above, and there IS an appeal process (endorser/co-signer or "extenuating circumstances" documentation).
- Best case: Get items DELETED (not just paid). Deletion removes the adverse history entirely.
- Backup: If PLUS is denied, Nik can appeal with documentation of resolved debts + an endorser.

**Action**: Prioritize deletions over payments. A charge-off that's deleted > a charge-off that's paid in full, for PLUS eligibility purposes.

---

## Does the June 2026 Decision Gate Still Hold?

**Yes, but the decision is now better-informed.**

By June 2026, Nik should know:
1. Actual scores after Phase 1-3 credit repair (target: 580-620+)
2. Building credit requirements from the 4 positive responders
3. Whether summer FAFSA aid is approved
4. Whether any job interviews materialized from the 454 applications

**The hybrid Path A-dominant recommendation still holds** but is now STRONGER because:
- The cheapest items (Verizon $207, CBE $368) are on credit reports and will actually move scores
- The FAFSA refund may already be in hand (check the Feb $9,454 anomaly)
- With $575 in Phase 1, Nik can remove 4 negative items this month alone
- The charge-off strategy is clearer: settle for 40-60% where possible, push for deletion

**Revised Path A/B split**:
- **Path A (credit repair)**: $3,711-$6,500 over 90 days — this is the primary allocation
- **Path B (deposit savings)**: Hold $3,000-$5,000 liquid as deposit reserve while repairing credit
- **Both are affordable** with $4,930/month surplus. There is no need to choose one path exclusively.

---

## Immediate Action Items (Next 7 Days)

1. **TODAY**: Call Verizon collections — negotiate pay-for-delete for $207. Get agreement in writing (email/letter) before paying.

2. **TODAY**: Call CBE Group (1-800-274-1908) — identify the original creditor for the $368 collection. Ask about pay-for-delete.

3. **THIS WEEK**: Reply to the 4 positive building responses — ask each:
   - "What credit score do you require for approval?"
   - "Which credit bureau do you pull?"
   - "Would a larger security deposit (2x-3x) offset a lower credit score?"

4. **APR 5**: If Diversified Adjustment Bureau has not responded to debt validation letter, dispute Verizon with all 3 bureaus citing failure to validate.

5. **APR 1-7**: Call Capital One (800-258-9319) — explore settlement options. Ask: "What would you accept to settle this account and request deletion from all credit bureaus?"

6. **APR 15**: Send WebBank/OM goodwill letter requesting late payment removal from TransUnion.

7. **CHECK NOW**: Was the Feb 2026 $9,454.54 deposit the spring FAFSA refund? If yes, that capital is already in hand. If no, check USD portal for refund status — it should be imminent.

---

## Summary

| Metric | Previous (Incorrect) | Corrected |
|--------|---------------------|-----------|
| #1 priority | TrueAccord $293 (not on reports) | Verizon $207 (all 3 bureaus) |
| Total negative balance | Unknown | $9,232 across 6 accounts |
| Cheapest action | $293 (zero impact) | $207 (3 bureau items removed) |
| Repair budget needed | Unknown | $3,711-$6,500 realistic |
| Target score by June | Vague | 580-620 (conservative) |
| Path recommendation | Hybrid A-dominant | Hybrid A-dominant (confirmed, now with real numbers) |

The strategy is cleaner now. Six real targets, prioritized by cost-effectiveness, with a realistic $5,500 budget to resolve most or all of them within 90 days. Verizon and CBE are the immediate wins. The three charge-offs are the heavy lift. WebBank is the outlier requiring a different approach.

Execute Phase 1 this week. $575 to remove 4 negative items is the highest-ROI action available.


---

# /root/STRATEGY_CREATIVE.md

# Creative Strategy — Non-Obvious Paths to First STR Unit
Last updated: 2026-03-28
Author: Opus (creative brainstorm session)

---

## Context
Standard paths (credit repair then apply, or save big deposit then apply) are slow or blocked by luxury building hard floors. This document explores angles Nikita has NOT yet considered.

---

## TIER 1 — HIGH FEASIBILITY, FAST TIME-TO-IMPACT

### 1. Father / Envitae as Leaseholder (Nikita as Operator)

**The idea**: Dad (or Envitae LLC) signs the lease. Nikita operates the STR. Nikita's credit never enters the picture.

**Why this might work**:
- Dad already has a relationship with these buildings — they know him, trust him, let him run STR operations
- Many buildings allow corporate/entity leases, especially for operators they already work with
- If Envitae already has units in the building, adding one more under the same entity is routine
- Dad's credit and income are what get checked, not Nikita's
- Nikita handles day-to-day operations, cleaning, guest management — a legitimate business arrangement

**Structure options**:
- **Sub-operator agreement**: Envitae leases the unit, Nikita manages it under an Envitae sub-contract, revenue splits agreed in writing
- **Management fee**: Nikita gets paid a management fee (20-30% of gross) + covers operating costs. This is exactly how property management works.
- **Profit share**: 50/50 after costs, with Nikita covering the deposit and Dad covering the lease qualification

**Why Nikita may have overlooked this**: It feels like "asking Dad for help" rather than doing it independently. But this is literally how property management businesses work. Nikita would be a legitimate operator, not a dependent.

**Feasibility**: 9/10 — depends entirely on Dad's willingness and credit
**Time to impact**: 2-4 weeks (lease signing speed)
**Risk**: Relationship risk if the unit underperforms. Mitigate with written agreement.

---

### 2. Authorized User Tradeline (Instant Score Boost)

**The idea**: Get added as an authorized user on someone's old, high-limit, perfect-payment credit card. The entire history of that card appears on Nikita's credit report.

**Why this works**:
- FICO scores include authorized user tradelines
- A card with 5+ years of history, $10K+ limit, 0% utilization, and perfect payments can add 30-80 points
- The person doesn't even need to give Nikita the card — just add him to the account
- Effect appears on credit report within 1-2 billing cycles (30-60 days)

**Who to ask**:
- **Dad** — if he has any old credit cards with clean history, this is the fastest single action
- **Family member or trusted friend** with excellent credit
- **Tradeline companies** — you can buy authorized user spots ($200-$800 per tradeline). Companies like BoostCredit101, Superior Tradelines, etc. This is legal but in a gray area — some lenders specifically exclude AU tradelines from underwriting, but FICO scores still include them

**Combined with Phase 1 credit repair ($575)**:
- Remove Verizon + CBE: +20-40 points
- Add 1-2 authorized user tradelines: +30-60 points
- Net effect: potentially +50-100 points in 60 days
- That takes 510-548 to 560-648 range — possibly above building thresholds

**Feasibility**: 8/10 (if Dad or family cooperates), 6/10 (if buying tradelines)
**Time to impact**: 30-60 days
**Cost**: $0 (family) or $200-$800 (purchased tradelines)

---

### 3. Guarantor / Co-Signer Service

**The idea**: Use a third-party guarantor service that co-signs your lease for a fee. These companies exist specifically for people with bad credit who can afford the rent.

**Real companies that do this**:
- **Insurent** (insurent.com) — acts as institutional guarantor. Requires proof of income (bank statements showing 27x monthly rent in annual income OR 50x in liquid assets). Fee: approximately 1 month's rent.
- **TheGuarantor** (theguarantor.com) — similar model, lower income requirements. Fee: approximately 65-90% of one month's rent.
- **Rhino** (sayrhino.com) — replaces security deposits with insurance. Doesn't help with credit check itself but reduces cash needed.

**Why this could work**:
- Nikita has $6,400/month provable payroll income from Alertify
- If rent is $1,800-$2,500/month, annual income of $76,800 meets the 27x threshold for most guarantor services
- The guarantor absorbs the credit risk — building gets their guarantee regardless of Nikita's score
- Many luxury buildings in major cities already accept Insurent/TheGuarantor — they're built for exactly this market

**Cost**: $1,800-$2,500 (one month's rent as fee) — but this is a one-time cost to get in NOW rather than waiting 3-6 months

**Feasibility**: 8/10 — depends on building accepting guarantor services
**Time to impact**: 1-2 weeks once approved
**Key question**: Do the 4 positive buildings accept institutional guarantors? Add this to the follow-up questions.

---

### 4. Non-Luxury First Unit to Bootstrap Cash Flow

**The idea**: Skip the luxury Envitae buildings for now. Get ANY unit that's STR-legal in a building with lax credit requirements. Use the cash flow to fix credit, then graduate to the target buildings.

**Where to look**:
- **Individual landlords** (not corporate buildings) — they often don't run credit checks or are flexible with deposits. Craigslist, Zillow rentals, local Facebook groups
- **Condo owners** who allow subleasing — individual condo owners can approve tenants themselves. Many are desperate for reliable tenants and care more about income proof than credit
- **Small multifamily buildings** (2-4 units) — mom-and-pop landlords are the most flexible
- **Lease takeovers / sublets** — someone breaking their lease may accept any creditworthy subletter, and the building's credit check may not apply to subletters
- **Extended stay / corporate housing platforms** — Furnished Finder, Landing — some operate more like Airbnb and less like traditional leases

**The math**:
- Cheap 1BR in a non-luxury building: $1,200-$1,500/month
- STR revenue for a decent unit: $2,500-$4,000/month
- Net profit: $1,000-$2,500/month
- This profit accelerates credit repair AND builds STR operational experience

**Why this is underrated**:
- Operational experience in a cheaper unit makes the luxury building pitch stronger
- Revenue from unit #1 funds credit repair to qualify for unit #2
- Even 3-4 months of operation creates a track record (reviews, revenue data) that's useful when approaching luxury buildings

**Feasibility**: 7/10
**Time to impact**: 2-4 weeks to find a unit, 1 month to start STR operations
**Risk**: Some cities have STR regulations — verify local ordinances

---

## TIER 2 — MODERATE FEASIBILITY, CREATIVE ANGLES

### 5. Master Lease Arrangement via Envitae

**The idea**: Propose a master lease to a building where Envitae already operates. Instead of a standard residential lease, Envitae (or a new LLC with Dad as member) signs a master lease for multiple units, with Nikita managing them.

**Why this is different from idea #1**:
- A master lease is a commercial arrangement, not a residential one
- Commercial leases have different underwriting — they look at business financials, not personal credit
- If Envitae already has 5-10 units in a building, proposing to master-lease 5 more at a guaranteed rate is attractive to building management
- Nikita becomes the operations manager for this block of units

**How to pitch it**: "Envitae wants to expand our presence in your building. We'd like to master-lease [X] additional units at [rate]. We'll handle all furnishing, guest management, and cleaning. Here's our track record in your building..."

**Feasibility**: 6/10 — requires Dad's active participation and building willingness
**Time to impact**: 4-8 weeks (commercial negotiations take longer)

---

### 6. Rapid Secured Credit Card Strategy

**The idea**: Open 2-3 secured credit cards NOW with small deposits ($200-$500 each). Use them for small purchases, pay in full each month. Combined with negative item removal, this builds positive tradelines fast.

**Why this stacks**:
- Removing negatives takes away points deductions
- Adding positive tradelines adds points
- Both together create a compounding effect
- After 3-6 months, secured cards often graduate to unsecured (returning the deposit)

**Best secured cards for credit building**:
- **Discover it Secured** — reports to all 3 bureaus, graduates automatically
- **Capital One Secured** — $49-$200 deposit, reports to all 3
- **OpenSky Secured** — no credit check to open, reports to all 3
- **Chime Secured Credit Builder** — no deposit needed, uses Chime checking balance

**Timeline**: Apply today, cards arrive in 7-10 days, first statement reports in 30 days, score impact in 60 days

**This is NOT a standalone strategy** but it stacks with everything else. +10-20 points from new positive tradelines on top of everything else.

**Feasibility**: 9/10 — anyone can get a secured card
**Time to impact**: 60-90 days for score impact
**Cost**: $400-$1,000 in refundable deposits

---

### 7. Experian Boost + UltraFICO + Rent Reporting

**The idea**: Stack every free/cheap score-boosting service available:

- **Experian Boost** (free) — adds utility and streaming payments to Experian report. Typical boost: +5-15 points. Nikita's FPL electric ($170/mo) and any streaming services qualify.
- **UltraFICO** (free) — uses bank account history (BofA checking with consistent balance) to improve score. Works well for people with thin/damaged credit but good banking behavior.
- **Rent reporting services** ($5-10/month) — services like Boom, RentTrack, or LevelCredit report your rent payments to credit bureaus. Nikita's $1,650/month rent becomes a positive tradeline.
- **Self.inc** ($25-$100/month) — credit-builder loan that reports to all 3 bureaus as an installment loan (adds loan diversity to credit mix)

**Stacking all of these**:
- Experian Boost: +5-15 points (Experian only)
- Rent reporting: +10-20 points (varies by bureau)
- Secured cards: +10-20 points
- Authorized user: +30-60 points
- Negative item removal (Phase 1): +20-40 points
- **Combined potential**: +75-155 points over 60-90 days

Even at the conservative end (+75), that takes 510 to 585 and 548 to 623. The 623 on Experian may clear building thresholds.

**Feasibility**: 9/10
**Time to impact**: 30-90 days (layered)
**Cost**: ~$50-$100 total

---

### 8. Negotiate an "Operator Exception" Directly with Building Management

**The idea**: Don't apply as a regular tenant. Approach the building as a BUSINESS PARTNER through Envitae.

**The pitch**: "My father's company, Envitae, operates [X] successful STR units in your building. We'd like to expand. I'll be the dedicated on-site operations manager for our units. Here's our revenue data, guest reviews, and operational track record. We're proposing to add a unit with me as the named tenant and operator."

**Why this reframes everything**:
- Regular tenants are evaluated on credit + income to assess rent payment risk
- A business partner / operator is evaluated on business performance and relationship
- If Envitae is already profitable in their building, the building management knows the rent will be paid
- This is a conversation with the building's operations team or ownership, NOT the leasing office
- The leasing office has rigid criteria; ownership has discretion

**How to execute**:
- Dad introduces Nikita to building management (not leasing agents — the actual decision-makers)
- Present: Envitae's track record, revenue data, guest satisfaction scores, insurance documentation
- Propose: Nikita signs lease as individual with Envitae as guarantor (Dad's company)
- Ask for: Exception to standard credit requirements based on business relationship

**Feasibility**: 7/10 — depends on Dad's relationships and building ownership structure
**Time to impact**: 2-6 weeks

---

### 9. LLC + Business Credit (Bypass Personal Credit Entirely)

**The idea**: Form an LLC, build business credit rapidly, lease the unit under the LLC.

**The reality check**: This is slower than people think. Business credit takes 3-6 months minimum to be meaningful. However, if Nikita starts NOW, it's ready by the June decision gate.

**Steps**:
1. Form LLC in Florida ($125 filing fee) — can do online today
2. Get EIN from IRS (free, instant online)
3. Open business bank account (Chase, BofA, etc.)
4. Get a DUNS number from Dun & Bradstreet (free)
5. Open business credit accounts that report to business bureaus: Uline, Grainger, Quill — these are "starter" net-30 accounts
6. Apply for business credit card (secured if needed)
7. After 3-4 months with 5+ reporting accounts, you have a Paydex score

**Problem**: Most residential buildings still require personal guarantees. BUT some buildings that work with corporate tenants (like buildings that already host Envitae units) may accept an LLC lease with business financials.

**Better use**: The LLC isn't for the lease credit check — it's for:
- Separating STR income (makes tax deductions cleaner)
- Professional appearance when approaching buildings
- Building long-term business credit for future units
- If combined with Dad as LLC co-member, his credit may be usable

**Feasibility**: 5/10 for bypassing personal credit, 9/10 as infrastructure for scaling
**Time to impact**: 3-6 months for meaningful business credit
**Cost**: $125 + ~$500 in starter account deposits

---

## TIER 3 — CREATIVE LONGSHOTS

### 10. Airbnb Co-Hosting (No Lease Needed)

**The idea**: Skip the lease entirely. Become a co-host on someone else's Airbnb listing. Zero credit check, zero deposit.

**How it works**:
- Find Airbnb hosts who are too busy to manage their listings
- Offer to manage their listing for a revenue split (20-40% of bookings)
- Airbnb has a built-in "co-host" feature — no lease, no liability, just operations
- Start earning STR income immediately while building credit and saving for your own unit

**Where to find hosts who need help**:
- Airbnb co-host marketplace (Airbnb recently launched this)
- Local Facebook groups for STR hosts
- Dad's Envitae network — other operators who are overwhelmed
- Turno, Hospitable, or other STR management platforms have job boards

**The bootstrap math**:
- Manage 3-5 units as co-host at 25% revenue share
- Average unit does $2,500/month, your cut: $625/unit
- 4 units = $2,500/month in STR income with zero capital deployed
- After 3-6 months: operational track record + saved cash + repaired credit, then your own unit

**Feasibility**: 8/10
**Time to impact**: 1-2 weeks to start, income within first month
**Cost**: $0

---

### 11. House Hack — Rent a 2BR, STR the Second Room

**The idea**: Move to a 2BR apartment with lenient credit requirements. Live in one room, Airbnb the other. This requires only a residential lease (not an STR-specific one) and many buildings allow room rentals on Airbnb.

**Why this is smart**:
- A 2BR might be $1,800-$2,200/month
- STR income from one room: $1,200-$2,000/month (in a desirable area)
- Net cost to Nikita: $0-$600/month for housing
- Current rent is ~$1,650, so this could be rent-neutral while generating STR experience
- Individual landlords for 2BR units are often more flexible on credit

**Feasibility**: 6/10
**Time to impact**: 2-4 weeks
**Risk**: Some leases prohibit subletting — must check terms

---

### 12. Private Money / Hard Money Partnership

**The idea**: Instead of renting, find a cheap condo to purchase using hard money or private money (from Dad's network). STR the condo. No landlord credit check — you own it.

**Reality check**: This requires significant capital ($20K-$50K minimum) and is a bigger play. But worth noting because:
- Hard money lenders care about the deal, not personal credit
- If Dad co-signs or provides down payment, Nikita operates
- Condo HOAs may restrict STR — must check
- Probably not viable now, but could be the "unit #2" path after cash flow from unit #1

**Feasibility**: 3/10 (now), 7/10 (with 6-12 months of STR cash flow)
**Time to impact**: 2-4 months if capital available
**Cost**: $20K-$50K

---

## RECOMMENDED COMBO PLAY (The Fastest Realistic Path)

Run these in parallel over the next 30-60 days:

### Immediate (This Week)
| Action | Cost | Impact |
|--------|------|--------|
| Phase 1 credit repair (Verizon + CBE) | $575 | -4 negative items |
| Sign up for Experian Boost | $0 | +5-15 points (Experian) |
| Sign up for rent reporting service | $10/mo | +10-20 points over 60 days |
| Apply for 2 secured credit cards | $400-$600 deposit | +10-20 points over 60 days |
| Ask Dad about authorized user tradeline | $0 | +30-60 points in 30-60 days |
| Start co-hosting on Airbnb (idea #10) | $0 | Immediate STR income + experience |

### This Month
| Action | Cost | Impact |
|--------|------|--------|
| Ask 4 positive buildings about guarantor services | $0 | Determines if Insurent/TheGuarantor is viable |
| Explore Envitae lease arrangement with Dad (idea #1) | $0 | Could bypass credit entirely |
| Pitch "operator exception" to buildings (idea #8) | $0 | Reframes from tenant to business partner |
| Form LLC in Florida | $125 | Infrastructure for professional STR operations |

### By June Decision Gate
| Metric | Conservative | Optimistic |
|--------|-------------|-----------|
| Credit score | 580-620 | 620-670 |
| STR operational experience | 2-4 months co-hosting | Active unit via Envitae arrangement |
| Liquid savings | $12K-$18K | $15K-$22K |
| Job status | Still searching + Alertify | New offer in hand |

---

## The Single Highest-Leverage Move

~~Talk to Dad about idea #1 or #8.~~ **VOIDED — Dad has been asked multiple times and is unwilling to cooperate. Envitae/Dad paths are off the table. See ideas #3, #4, #10 instead.**

**Real highest-leverage move:** Guarantor service (Insurent/TheGuarantor) + Airbnb co-hosting to build track record while credit repairs. Everything else is optimization. If Envitae can add a unit with Nikita as operator — bypassing the personal credit check entirely — the entire credit repair timeline becomes irrelevant for the first unit. Credit repair still matters for scaling (units 2, 3, 4 under Nikita's own name) and for Grad PLUS protection, so continue Phase 1-4 regardless.

But the fastest path to "first unit live and cash-flowing" is almost certainly through the Envitae relationship, not through the front door of a leasing office.

---

## Questions to Ask Dad This Week

1. Can Envitae lease a new unit with me as the named operator?
2. Would you be willing to add me as an authorized user on one of your credit cards?
3. Which building managers do you have the strongest relationships with?
4. Can we pitch an "expansion" to a building where Envitae already operates?
5. What are the actual margins on existing Envitae units? (Need this for business case)

---

## Summary Table

| # | Idea | Feasibility | Time to Impact | Cost | Bypasses Credit? |
|---|------|-------------|----------------|------|-----------------|
| 1 | Dad/Envitae as leaseholder | 9/10 | 2-4 weeks | $0 (deposit TBD) | YES |
| 2 | Authorized user tradeline | 8/10 | 30-60 days | $0-$800 | No (improves score) |
| 3 | Guarantor service (Insurent) | 8/10 | 1-2 weeks | ~1 month rent | YES |
| 4 | Non-luxury first unit | 7/10 | 2-4 weeks | First/last/deposit | YES (easier buildings) |
| 5 | Master lease via Envitae | 6/10 | 4-8 weeks | $0 | YES |
| 6 | Secured credit cards | 9/10 | 60-90 days | $400-$1000 | No (improves score) |
| 7 | Experian Boost + rent reporting | 9/10 | 30-90 days | ~$50 | No (improves score) |
| 8 | Operator exception pitch | 7/10 | 2-6 weeks | $0 | YES |
| 9 | LLC + business credit | 5/10 | 3-6 months | $625+ | Partially |
| 10 | Co-hosting (no lease needed) | 8/10 | 1-2 weeks | $0 | YES |
| 11 | House hack 2BR | 6/10 | 2-4 weeks | Lease costs | No (still need lease) |
| 12 | Hard money purchase | 3/10 | 2-4 months | $20K-$50K | YES |


---

# /root/FAFSA_2026.md (financial / loan strategy)

# FAFSA & Student Loans — Nikita Rogers (2026)

Source: studentaid.gov + University of San Diego portal + ChatGPT analysis
Last updated: 2026-03-26

---

## Current Loan Balance (studentaid.gov, as of 03/01/2026)

**Total Balance: $100,656.47**
- Principal: $95,666.00
- Outstanding interest: $4,990.47

### Loan Breakdown

| Servicer | Balance | Rate | Type | Status |
|----------|---------|------|------|--------|
| Dept of Ed / Nelnet | $21,052.59 | 7.94% | Unsubsidized | In School |
| Dept of Ed / Nelnet | $20,821.40 | 8.08% | Unsubsidized | In School |
| Dept of Ed / Nelnet | $22,370.97 | 8.94% | Graduate PLUS | Deferment |
| UIUC (Perkins) | $1,099.00 | 5.00% | Perkins | Deferment |
| UIUC (Perkins) | $2,199.00 | 5.00% | Perkins | Deferment |
| SAVE Plan — Subsidized | $4,061.94 | 5.05% | Subsidized | Deferment |
| SAVE Plan — Subsidized | $5,575.05 | 4.45% | Subsidized | Deferment |
| SAVE Plan — Unsubsidized | $2,332.94 | 4.45% | Unsubsidized | Deferment |
| SAVE Plan — Subsidized | $5,563.46 | 3.76% | Subsidized | Deferment |
| SAVE Plan — Unsubsidized | $2,344.46 | 3.76% | Unsubsidized | Deferment |
| SAVE Plan — Subsidized | $4,559.20 | 4.29% | Subsidized | Deferment |
| SAVE Plan — Unsubsidized | $2,495.51 | 4.29% | Unsubsidized | Deferment |
| SAVE Plan — Subsidized | $3,550.02 | 4.66% | Subsidized | Deferment |
| SAVE Plan — Unsubsidized | $2,630.93 | 4.66% | Unsubsidized | Deferment |

Repayment plans active:
- SAVE (Income-Driven, recertification 12/22/2027): $340/mo — currently in deferment
- Standard (fixed): $415/mo — currently in deferment
- In-school deferment for current USD loans: $0/mo

---

## 2025–26 Aid Year — University of San Diego

### Cost of Attendance (Annual)
| Item | Amount |
|------|--------|
| Tuition | $11,580 |
| Slack Fee | $40 |
| **Total Billable** | **$11,620/semester** |
| Books & Supplies | $540 |
| Housing (off campus) | $15,219 |
| Food (off campus) | $7,533 |
| Personal Necessities | $5,067 |
| Transportation | $1,989 |
| Avg Grad PLUS Loan Fee | $1,460 |
| Avg Stafford Loan Fee | $244 |
| **Total Non-Billable** | **$32,052** |
| **Total COA** | **$43,672** |

### Loans Accepted (2025–26)
| Type | Fall 2025 | Spring 2026 | Annual Total | Status |
|------|-----------|-------------|--------------|--------|
| Fed Direct Unsubsidized | $10,250 | $10,250 | $20,500 | ACCEPTED |
| Fed Direct Grad PLUS | $10,856 | $10,856 | $21,712 | ACCEPTED |
| **Total** | **$21,106** | **$21,106** | **$42,212** | |

### Cash Refund Calculation
- Loans per semester: $21,106
- Tuition/fees per semester: $11,620
- **Refund per semester: ~$9,486**
- **Annual cash to Nikita: ~$18,972**

School pays tuition directly; remaining loan balance is refunded to Nikita for housing/living.

### Summer 2026
- Summer Financial Aid Application: submitted March 2026
- Status: **RECEIVED / NOT YET REVIEWED**
- Summer amount: unknown — school processes May/June 2026
- Check back: USD portal → Financial Aid Offers

---

## Federal Borrowing Limits & Remaining Capacity

| Limit Type | Cap | Used | Remaining |
|-----------|-----|------|-----------|
| Federal lifetime (grad + undergrad) | $138,500 | ~$100,656 | **~$37,844** |
| Grad PLUS | No hard cap | $22,371 so far | Unlimited (credit-check) |

**Remaining ~$37K in direct loans** = roughly 1.5–2 more semesters of Unsubsidized borrowing.
After that, all future borrowing is **Grad PLUS only** (requires clean credit — no delinquencies/charge-offs).

---

## Debt Trajectory

| Milestone | Projected Total Debt |
|-----------|---------------------|
| Today (Mar 2026) | $100,656 |
| After current USD program | ~$142,000 |
| If second master's (typical) | ~$200,000–$260,000 |

Monthly payment at $140K debt @ avg ~6.5% interest (standard 10yr):
→ ~$1,590/month

Monthly payment at $200K debt:
→ ~$2,270/month

Currently in deferment (in school) — payments begin 6 months after graduation.

---

## Strategic Notes

1. **The ~$9,500/semester refund is the only flexible capital from FAFSA.** Everything else flows directly to tuition. Each refund is a deployment decision.

2. **Grad PLUS eligibility depends on clean credit.** Any delinquency or charge-off can kill future Grad PLUS access. This makes credit repair directly tied to FAFSA borrowing capacity — they are not separate concerns.

3. **Don't think of FAFSA loans as free money.** At $100K+, interest is accruing even in deferment on unsubsidized/PLUS loans. Use refunds for income-generating purposes only.

4. **Summer aid is unconfirmed.** Do not count on it until school reviews (check portal May/June 2026).

5. **Second master's is possible but costly.** Direct loan room is nearly exhausted (~$37K left). A second master's = almost entirely Grad PLUS = ~$40K–$70K/year borrowed. Total debt could reach $200K+.

---

## Action Items
- [ ] Check USD portal (Financial Aid Offers) in May/June for summer 2026 amount
- [ ] Upload credit report PDFs to credit dashboard to understand Grad PLUS risk (delinquencies affect eligibility)
- [ ] Decide deployment of next refund (~$9.5K): Path A (credit payoff) vs Path B (security deposit)
- [ ] Track monthly burn rate via Finance dashboard (142 BofA transactions) to know true monthly need vs. surplus

---

## Source Links
- studentaid.gov/my-aid — current balances, loan details
- USD portal → Financial Aid Offers — upcoming disbursements
- /root/MASTER_PLAN.md — strategic context for how FAFSA fits the bigger picture


---

# Memory: feedback_apples_systems_demo.md

---
name: Apples AI Systems cards must have live demos
description: Only list a system in the apples.live "AI Systems" card if there's a working applicable demo (live page, phone number, or interactive tool). No claims without proof.
type: feedback
originSessionId: 46236397-205b-455d-9cc3-6ae389f85966
---
Every entry in the "AI Systems" section on apples.live (landing page) must point to a working live demonstration. No marketing claims about a capability unless a visitor can interact with it on the spot.

**Why:** Nik whittled the section from 11 cards down to 5 specifically because the cut ones lacked live demos. He flagged this as a guardrail — future additions/restorations must clear the same bar. "We just wanna make sure that when we put them back, we have a live system to demonstrate, we don't wanna just say something without having applicable live demonstration."

**How to apply:**
- Before adding/restoring a card, confirm there is a real demo URL or live phone number that works end-to-end (200 OK, real interaction, not just a screenshot or mock).
- If a past project was cool but has no demo, build the demo first, then add the card. Don't add the card and promise to wire a demo later.
- Demos for the current 5 (2026-04-25) live at: `/audit.html`, `tel:+18137338792`, `/workflow.html`, `/seo.html`, `/dashboard-demo.html` — all served by CallFlow (port 4089) proxied through apples.live and red.apples.name nginx vhosts.
- The 6 removed cards (whittled out) are not documented; if Nik wants them restored, ask which ones and verify each has a demo URL before adding.


---

# Memory: feedback_html_element_removal.md

---
name: Always grep JS before removing HTML elements
description: Removing an HTML element without checking JS references causes null crashes that break unrelated UI
type: feedback
---

Always grep for an element's ID in JS before removing any HTML element from the frontend.

**Why:** Removed `btn-voice` button from index.html without checking — `getElementById('btn-voice')` returned `null`, `.addEventListener` on null threw a TypeError, crashing the entire script and breaking the mobile toolbar and paste box. A completely unrelated part of the UI broke silently.

**How to apply:** Before removing any HTML element with an `id`, run `grep -n "btn-id-name"` across the JS/HTML file first. If references exist, either remove them too or guard with a null check (`if (el) { ... }`).


---

# Memory: feedback_loop_focus.md

---
name: Autoresearch loops - only trading
description: Nik wants autoresearch loops focused on trading only, not all 11 projects
type: feedback
---

Don't run autoresearch loops on every project. Focus on trading only.

**Why:** Running 11 loops simultaneously burned through Claude Max quota fast, most iterations failed from rate limiting (~39,000 wasted attempts). Spreading across all projects was too broad.

**How to apply:** When starting loops, only start `loop start trading`. Don't suggest starting loops for other projects unless Nik specifically asks.


---

# Memory: feedback_no_dashboards_for_config.md

---
name: No config dashboards - build invisible pipelines
description: Nik doesn't want to manually configure webhooks/triggers via dashboards. Build integrations that just work automatically.
type: feedback
originSessionId: 6033875d-5918-417c-9923-8bdc934309bc
---
Nik won't use dashboards to set up webhook routing or configure triggers. Doesn't see himself going to a dashboard and setting stuff up.

**Why:** He wants to do natural things (take meetings, get messages, browse jobs) and have agents handle the follow-up automatically. Config UIs are friction.

**How to apply:** When building automation, never make the user configure it through a UI. Pre-wire specific integrations that activate automatically. Dashboards are OK for monitoring (what happened, what's running) but not for setup. Conway's value is invisible plumbing, not another dashboard.

---

# Memory: feedback_screen_timeout.md

---
name: Extend phone screen timeout before long automation runs
description: Before multi-step phone automation, bump screen_off_timeout so display doesn't sleep mid-task
type: feedback
originSessionId: 46236397-205b-455d-9cc3-6ae389f85966
---
Before any multi-step phone automation, extend screen timeout to prevent mid-run sleep.

**Why:** Phone display sleeps during long tap/type/screenshot sequences, breaking automation partway through. Screencap/input still work on locked screen but WebView content may pause.

**How to apply:**
```
phone rsh 'settings put system screen_off_timeout 600000'   # 10min
# restore default when done:
phone rsh 'settings put system screen_off_timeout 30000'    # 30s
```
Run at start of any automation expected to exceed default timeout (30s–2min).


---

# Memory: feedback_text_field_clearing.md

---
name: Clear text fields with select-all + delete, not backspace loop
description: When clearing text on phone/web automation, use Ctrl+A (or triple-tap) + delete, never loop backspace per char
type: feedback
originSessionId: 46236397-205b-455d-9cc3-6ae389f85966
---
Never clear text by looping backspace one char at a time. Use select-all + delete.

**Why:** Nik called out a 40-char backspace loop as absurdly slow. One keystroke vs 40.

**How to apply:**
- Android via rish: `input keycombination 113 29` (Ctrl_Left + A), then `input keyevent 67` (Del)
- Or triple-tap to select word/line, then delete
- Applies to any text-clear op in phone/browser automation


---

# Memory: project_context_injection_architecture.md

---
name: Context Injection Architecture Spectrum
description: Nik's understanding of MD file count as a proxy for context injection complexity — autoresearch(1) vs his setup(2) vs OpenClaw(8)
type: project
---

Nik has mapped agent context injection as a spectrum measured by number of MD files:

- **autoresearch (karpathy)**: 1 MD (program.md) + 2 Python files + 1 TSV output = minimal, fully autonomous loop
- **Nik's VPS setup**: 2 MD files — GLOBAL.md (VPS-wide, all agents) + PROJECT.md (per-project, per-session). Clean global/local split.
- **OpenClaw**: 8 MD files — routing, agent personas, tool definitions, channel-specific configs, etc.

**Why:** Nik discovered this while building a custom autoresearch dashboard and a custom OpenClaw dashboard and comparing them. The insight is that MD file count = context injection surface area.

**How to apply:** Nik's 2-file architecture is intentionally minimal and correct for a single-user, multi-project setup. Don't suggest adding more MD files unless there's a concrete need (e.g., different agent personas or channel-specific routing). The missing piece vs autoresearch is a structured output/metric file (like autoresearch's TSV) that feeds back into agent context — this would close the loop for the Opus strategy agent.


---

# Memory: project_google_drive_quota.md

---
name: Google Drive quota full — can't create new Sheets/files
description: 2026-04-24 — Drive API returns storageQuotaExceeded on file create. Use SQLite + HTML dashboards instead of Google Sheets until quota is cleared.
type: project
originSessionId: 46236397-205b-455d-9cc3-6ae389f85966
---
Nik's Google Drive is full. API create returns `storageQuotaExceeded`.

**Why:** Confirmed 2026-04-24 trying to create "Marketing Posts Tracker". All new Drive file creation fails.

**How to apply:**
- Don't try to create new Google Sheets for trackers/dashboards. Use SQLite + HTML on VPS instead.
- Can still read/edit EXISTING sheets via `mcp__connectors__sheets_append` (append is in-place, no new storage).
- For spreadsheet-like UX, build sortable HTML table + CSV export endpoint. Nik can upload the CSV into Sheets once quota is free.
- Working refresh token + client secret combo: `GOOGLE_REFRESH_TOKEN_777` + `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` (Drive client ID).
- Reauth URL: https://45-55-169-94.nip.io:8443 (won't fix quota, only refresh).


---

# Memory: project_manage_tool.md

---
name: Manage tool for session orchestration
description: manage CLI replaces old autoloop -- Sonnet reads session output, crafts targeted next instruction each round instead of repeating same message
type: project
---

`manage` CLI at `/usr/local/bin/manage` replaces old `autoloop` system.

**Why:** Old autoloop sent same dumb message every round. New tool has Sonnet analyze session output and craft specific next instructions each round.

**How to apply:** Use `manage start <session> "goal" <rounds>` instead of autoloop. Use `manage once` for one-shot. Costs ~$0.01-0.03/round (Sonnet). Auto-detects goal completion and stops. Pings Nik on Slack when done.

Commands: `manage start|once|stop|status|log <session>`


---

# Memory: project_namecheap_api.md

---
name: Namecheap API access
description: Namecheap API works — username is nikitarogers (NOT Fantastic-Eye-672 which is Reddit), VPS IP whitelisted, key in keys333
type: reference
originSessionId: 46236397-205b-455d-9cc3-6ae389f85966
---
Namecheap API is fully working from VPS (45.55.169.94 whitelisted).

- Endpoint: `https://api.namecheap.com/xml.response`
- ApiUser / UserName: `nikitarogers` (fixed 2026-04-24 — keys333 previously had Fantastic-Eye-672, which is Reddit's username)
- ApiKey: in `/root/keys333` as `NAMECHEAP_API_KEY`
- ClientIp: `45.55.169.94`

Prefer API over browser/phone automation — Namecheap's login form has invisible CAPTCHA / bot detection that silently blocks automated logins. API has no such block.

Common commands used:
- `namecheap.domains.getList` — list all domains
- `namecheap.domains.dns.getHosts` — list DNS records
- `namecheap.domains.dns.setHosts` — REPLACES all DNS records (include EmailType=FWD to preserve email forwarding)
- `namecheap.domains.dns.getEmailForwarding` / `setEmailForwarding` — catch-all use `MailBox=*`

Domains owned (as of 2026-04-24): aiagentplaybook.org, apples.live, apples.name, askcraig.org, textcraig.org, youtubetranscript.us.


---

# Memory: project_per_client_platforms.md

---
name: Per-client SMB platforms (productizing Nik's tools)
description: New business concept — package the tools Nik already uses for each SMB client into a custom platform per business. Pilot client is Bob Rogers (ceramic mosaic art).
type: project
originSessionId: 939238b7-c36a-48dd-9ae7-e222e31537b7
---
**Concept:** Productize the workflows Nik runs by hand for each client into a per-business web platform. Each instance bundles the specific tools that client needs — not a generic SaaS, a custom-fit dashboard.

**Pilot: Bob Rogers (ceramic mosaic art).** His instance needs two features:

1. **Pool permit contact scraper → one-click outreach**
   - Scrape pool permit contacts (existing tool Nik runs)
   - Pre-filled email template (editable in platform)
   - One-click approve & send

2. **Multi-platform social posting (Instagram + others)**
   - Analyze prior posts (images + captions) to extract style/voice
   - Generate content schedule
   - Per slot, three paths: upload own photo + caption, OR pick from pre-made templates Nik creates, OR one-click approve a generated draft
   - Review + post directly from platform (no leaving to native apps)

**Why:** Tied to Intelra observation that the consultant model is shifting from "tools" to "full platforms custom for each business." Nik wants to experiment with that pattern using clients he already works with.

**Why: How to apply:** Treat this as the canonical pattern for future client work — bundle existing automations into a per-client UI rather than handing off scripts. Bob's platform is the template; each new client gets a fork with their specific tools wired in.

**Status (2026-04-27):** Concept stage. No code yet. Probable next step: scaffold `/root/projects/bob-platform/` with Postgres schema for permit contacts + post schedule, FastAPI backend, simple Next.js UI, Instagram Graph API for posting.


---

# Memory: project_phone_bridge.md

---
name: Phone bridge (Pixel 9a)
description: Nik's Pixel 9a has no lockscreen PIN — Claude can unlock it via swipe. Full remote control via /usr/local/bin/phone.
type: project
originSessionId: 46236397-205b-455d-9cc3-6ae389f85966
---
Nik's Pixel 9a has NO PIN / NO biometric lockscreen. Lockscreen is swipe-only.

**Why:** Nik confirmed directly ("there is no pin"). Enables full autonomous control.

**How to apply:** To unlock from VPS, just run:
```
phone key 224              # wake screen (KEYCODE_WAKEUP)
phone swipe 540 2000 540 400 300   # swipe up to unlock
```
No PIN entry needed. After that, tap/type/open/screenshot all work normally.

**Bridge:** `/usr/local/bin/phone` — subcommands: sh, rsh, screenshot, tap, swipe, text, key, open, a11y, pkgs, top, notify, toast, clip, battery, location, photo, sms, status.

**Tunnel:** reverse SSH phone→VPS on port 22022. Autossh watchdog on phone. VPS monitor (systemd timer `phone-monitor.timer`) Slacks if dead >5min. Termux:Boot installed — auto-reconnects after phone reboots.

**Persistence layers:**
1. `~/.termux/boot/99-tunnel` — auto-run on boot via Termux:Boot app (now also calls `termux-wake-lock` as of 2026-04-26)
2. `~/watchdog.sh` — restarts autossh every 30s if dead
3. Shizuku (rish at /data/data/com.termux/files/usr/bin/rish) must be re-STARTED (open Shizuku app → Start) after phone reboot. Pairing persists, service does not. No workaround without root.

**Full setup doc:** `/opt/phone-automation.md` — verified working 2026-04-26. Includes phone-side settings (Termux:Boot, Tailscale Always-on VPN, battery whitelist), recovery flows, common automation patterns, and rooting trade-offs.

**Lockscreen wake gotcha:** On Android, swipe-up from lock sometimes opens notification shade instead. If that happens, `phone key 4` (back) then retry swipe.


---

# Memory: project_red_apples.md

---
name: Apples (formerly Red Apples) — customer-facing, freeze
description: Apples is Nik's B2B consulting + outreach site. Renamed from Red Apples on 2026-04-24 with domain switch to apples.live. Codebase NOT at /root/projects/matt/ — actual app is at /root/projects/LLM/web/.
type: project
originSessionId: 46236397-205b-455d-9cc3-6ae389f85966
---
**Brand rename (2026-04-24):** "Red Apples" → "Apples". New domain `apples.live` (old `red.apples.name` still resolves, points to same backend).

**Codebase location:** `/root/projects/LLM/web/` (Next.js, runs on :4094). NOT `/root/projects/matt/` — that's a different/older codebase, frozen and abandoned. The `smatt` tmux session was just where Nik worked on the LLM/web app, not where the code lives.

**Backend services:** ports 4089, 4094, 4104. Nginx configs: `/etc/nginx/sites-enabled/red-apples` and `/etc/nginx/sites-enabled/apples-live` (both proxy to :4094).

**Live data:**
- DB: `/root/projects/LLM/web/data/agentaudit.db` (SQLite, drizzle)
- Tables include: `users`, `clients`, `outreach_packs`, `audit_reports`, `client_milestones`, `outreach_prospects`
- Real customers: Bob Rogers (Ceramic Mosaic Art LLC), Anna Cordin (Gala Ballroom), Josh Envitae, Matt Smith
- Outreach packs: pre-auth links at /pack/{token} let consultants work without signing up

**Why:** Prospects and paying customers visit this site. Visual or functional regressions hurt revenue and trust. Real auditing/CRM data lives in agentaudit.db.

**How to apply:**
- Never bundle into internal UI refactors (shared components, theme changes, etc).
- Any change requires explicit Nik approval + staged rollout + visual check.
- When working on the codebase, work in `/root/projects/LLM/web/` (NOT `/root/projects/matt/`).
- The internal ops console at port 4120 (`/opt/red-apples/dashboard/server.py`) READS from agentaudit.db for the Consultants and Clients tabs — read-only, safe.


---

# Memory: project_terminal_state.md

---
name: Terminal — Confirmed Working State (2026-03-30)
description: All scroll fixes, CLI wrappers, rendering settings, and projects dashboard layout confirmed working as of 2026-03-30. Do not undo these.
type: project
---

Nik is very happy with the current terminal state. Do not revert any of these.

## Scroll System (`/opt/tty/public/index.html`)

**Momentum physics:** friction `0.95`, cutoff `0.005` — feels good, no slow tail.

**Rolling velocity window:** velocity computed from last 100ms of touchmove samples (not just last sample) — prevents wrong-direction momentum from finger jitter at end of gesture.

**Absolute target position:** momentum tracks its own `target` variable and writes `vp.scrollTop = target` each frame, so xterm interference between frames can't derail momentum.

**iOS snap-back fix:** `touchstart` is `passive:false`. When `wasMomentumActive` (momentum was running when finger touched), calls `e.preventDefault()` — prevents iOS Safari from resetting scroll to its last "committed" position on the stop-tap. This was the hardest bug to find.

**Momentum stop → no keyboard:** `wasMomentumActive` flag suppresses keyboard open when the tap was just to stop momentum (not a real tap).

**xterm scroll restore:** `userScrolled` + `savedScrollTop` + `restoreRafId` system restores scroll position after xterm writes. `cancelMomentum()` also cancels `restoreRafId` so orphaned rAFs can't fire after a stop-tap.

**Native scroll disabled:** `vp.style.touchAction = 'none'` + `touchmove` is `passive:false` with `e.preventDefault()` — prevents browser's own momentum from running in parallel with ours.

**Side scroll track width:** `width:50%` on both sides — meet in the middle, auto-adapts to any screen width.

## Global Terminal Environment (`/opt/tty/server.js`)

`TERM=dumb`, `COLUMNS=220`, `LINES=50` — COLORTERM removed. Applies to all web terminal sessions. Nik already has simple/screen-reader mode on in Claude Code and Gemini, so dumb term is consistent.

## CLI Wrappers

- `/usr/local/bin/gemini` — wraps `/usr/bin/gemini` with `TERM=dumb` (rendering fix)
- `/usr/local/bin/codex` — wraps `/snap/bin/codex` with `--no-alt-screen` (inline mode, no fullscreen TUI, colors preserved)

Both take precedence over system binaries via PATH (`/usr/local/bin` before `/usr/bin` and `/snap/bin`).

## Floating scroll button (`/opt/tty/public/index.html`)

`52×52px`, `background:transparent`, `border:none`, `color:#888` — no circle/shadow. `scrollToBottom()` uses 4 RAF passes + 120ms safety net to handle xterm lazy rendering.

## Port 4025 Projects Dashboard (`/opt/projects-api.js`)

Session chips use `target="_blank"` — fixes sessions not opening on mobile.

**No card titles** — `.card-name`, `.card-header`, `.global-name` all removed. Cards start with session chips directly.

**No header bar** — floating `+` (add project) and notification number at `position:fixed; top:8px; left:50%`. `+` centered, number to its right.

**padding-top: 52px on body** — cards clear the floating icons at `scrollTop=0`. No `scrollTo()` on load (causes overlap).

**Notification badge** — inline 14px monospace text; green `0` when all read, yellow number when unreads.

**`+` and `*` buttons** — white `#f0f0f0`.

**Card grid** — `align-items: start`, `.md-preview` `max-height: 140px`. Do not stretch cards.

**Global card** — identified by `# GlobalMD` heading at top of `/root/GLOBAL.md`, renders in MD preview. No separate title needed.


---

# Apples nikita-rogers wiki/analysis.md

## Overview

Nikita Rogers (GitHub: nikitarogers333) is a solo developer building apples.live, an AI-powered operations platform targeting small and micro businesses that are too small for enterprise tools like ServiceTitan or Toast. The product bundles roughly 18 modules: AI voice receptionist, booking, dispatch, quoting, payments, CRM, review management, lead generation, SEO, content creation, workflow automation, meeting transcription, a knowledge base, inventory, finance, and a headless browser agent. Six vertical-specific dashboards exist for trades, restaurants, outdoor services, retail, healthcare, and personal care. The platform is very early stage: apples.live was registered only on 2026-04-24, the demo is a single branded mock-up ("Bright Spark Electric"), and no named customers or pricing are publicly visible. A secondary product, youtubetranscript.us, is already priced ($9 to $79/month tiers) and shares the same SMS contact number, suggesting it predates apples.live or serves as a proving ground.

The founder appears to be a current or recent University of San Diego data science student operating out of Tampa, FL (lead generation demo data is Tampa electrical permits). The 917 area code on the shared phone number suggests prior New York ties. No personal LinkedIn, Crunchbase, About page, or press coverage exists under this name for this product, making brand visibility effectively zero outside owned properties. The web crawl also surfaced a Nikita Rogers who is a Northern Ireland fiction author and tattoo artist with a separate website, Goodreads profile, and paranormal novel series; this person appears to be an unrelated name collision and has no connection to apples.live.

The most distinctive product feature is permit-based lead generation: the sample leads CSV shows homeowner names, addresses, phone numbers, emails, permit types, and estimated job values pulled from county GIS and permit portals in Tampa within the past week. For electricians and contractors, this is a high-intent outreach list that does not require any ad spend. The demo metrics advertised (94% missed-call recovery, $3.20 cost per lead, 1.2-second response time) are compelling if real, but all data visible is synthetic and no production evidence exists yet.

## Observations

- The product demo is entirely wrapped inside a single "Bright Spark Electric" electrician persona, with no vertical switching or multi-tenant demo visible, suggesting the electrician/trades wedge may be the actual first real deployment target.
- All dashboard tabs (booking, CRM, leads, payments) render identical content, indicating the multi-tab architecture is visual scaffolding rather than functional differentiation at this stage.
- Lead generation via county permit scraping is technically sophisticated and legally grey; it surfaces contact details for people who filed permits but did not solicit contact, which may face CAN-SPAM or state-level cold-outreach scrutiny.
- The sample leads CSV is publicly accessible at apples.live/sample-leads.csv, which exposes real homeowner names, phone numbers, and email addresses scraped from Tampa public records without any authentication gate.
- apples.live was registered six days before the audit date, making this an extremely early-stage launch with essentially no SEO age or backlink history.
- The Namecheap nameservers (not Cloudflare) on apples.live suggest infrastructure is not yet hardened or CDN-optimized.
- youtubetranscript.us shares the same phone number as apples.live; if it has paying customers, it is the only confirmed revenue signal in the entire portfolio.
- Sixty-plus third-party integrations are listed on the homepage (Stripe, QuickBooks, HubSpot, Vapi, ElevenLabs, Twilio, n8n, etc.), but no integration actually appears wired in the demo, so depth of implementation is unknown.
- The GitHub repos (ai-agent-system, tmux-dash, terminal) plus a forked USD cloud-computing course suggest the founder is building the AI stack from scratch rather than wrapping no-code tools.
- No pricing, no trial, and no sign-up flow are visible on apples.live, meaning the product is not self-serve yet and every customer requires a direct sales conversation.
- The competing brand surface (red.apples.name) is not represented in this crawl, leaving its current state and relationship to apples.live unclear.
- HIPAA-compliant healthcare and EMR integration is claimed as a vertical feature but would require certification and legal review that a solo student-stage founder has almost certainly not completed.
- The askcraig.org domain (registered 2026-04-10) is not publicly resolvable; its purpose is unknown but the name suggests a separate AI assistant product in early testing.


---

# Apples nikita-rogers wiki/overview.md

---
type: overview
title: Nikita Rogers
updated: 2026-04-30
sources_count: 1
entity_count: 9
concept_count: 1
---

# Nikita Rogers

Solo operator building an AI tools platform (apples.live / red.apples.name) targeting small businesses across six verticals (residential trades, outdoor services, healthcare, restaurants, retail, personal care), bundling approximately 16-18 tools including voice receptionist, booking, CRM, dispatch, payments, and workflow automation. Also operates a separate SaaS product (youtubetranscript.us, YouTube transcript API, $9-$79/mo tiers). A recently registered domain (askcraig.org, 2026-04-10) exists with no public content. GitHub profile (nikitarogers333) includes repos ai-agent-system, tmux-dash, and terminal, plus a forked University of San Diego Data Science class repo. [src:./raw/audits/audit-2026-04-30.md]

## What's in this wiki

- 9 entities ([[index]] for the full list)
- 1 concept
- 1 ingested source

## Recent additions

- 2026-04-30: full rebuild from /raw/audits/audit-2026-04-30.md — created all entity/concept/source pages from scratch


---

# Apples nikita-rogers wiki/summary.md

## Apples / apples.live

Nikita Rogers is a solo founder building apples.live, an AI-powered operations platform aimed at small businesses in trades and service verticals. The domain was registered on April 24, 2026 via Namecheap, making it a very new property. A second surface, red.apples.name, appears to present the same or closely overlapping product. Neither site publishes a founder bio, location, About page, or social links.

The platform bundles approximately 18 tools into a single dashboard: an AI voice receptionist, booking and scheduling, quoting, payments (Stripe, ACH, QuickBooks sync), CRM, dispatch, inventory, lead generation, SEO content writing, review management, workflow automation, a headless browser agent for vendor portal scraping, meeting transcription, a knowledge base, and financial document extraction. The homepage positions these as purpose-built suites for six verticals: trades (HVAC, electrical, plumbing), restaurants, outdoor services (lawn, pool, pest), retail showrooms, healthcare practices, and personal-care studios. The trades framing explicitly targets shops too small for ServiceTitan or Jobber.

## Demo: Bright Spark Electric

The live product demo on the site is branded as "Bright Spark Electric," a fictional electrical contractor. The demo dashboard shows an AI agent handling inbound calls, qualifying leads, and booking appointments. Illustrative metrics in the demo include 23 calls processed in a day, a 2% missed-call rate, 1.2-second average response time, 62% booking rate, $3.20 cost per lead, and $58,400 in pipeline across 25 active leads. The AI voice agent in the demo is named Jamie and responds to questions about pricing, qualification, and scheduling. Scripts are versioned (v3.2 in the demo). Calendar integration points to Google Calendar; CRM integration points to HubSpot.

The demo script shows the receptionist asking four qualification questions -- emergency or scheduled, nature of the issue, address, and availability -- then booking the job and sending an SMS confirmation. After-hours calls go to an AI voicemail that queues a callback. Three voice personas are available: Jamie (warm, female), Marcus (calm, male), and Riley (energetic, neutral).

## Lead Generation Tool

One component of the platform surfaces leads sourced from public building permit databases. A sample CSV exposed on the site contains Tampa, Florida electrical permit records from late April 2026 -- homeowners who filed for panel upgrades, EV charger installs, solar tie-ins, rewires, and generators -- with names, addresses, phone numbers, emails, permit numbers, filed dates, and estimated job values ranging from $2,400 to $11,200. This suggests the leads module pulls from county GIS and permit portal feeds and delivers enriched, high-intent prospects to trades contractors.

## Integrations

The platform advertises connections to a wide range of third-party tools: Anthropic, OpenAI, Vapi, ElevenLabs, and Deepgram on the AI and voice side; Stripe, Square, QuickBooks, FreshBooks, Wave, and Xero for payments and accounting; HubSpot, Salesforce, Pipedrive, and Zoho CRM; Twilio, SendGrid, and Postmark for messaging; Google Workspace, Microsoft 365, Slack, and Notion for productivity; Shopify, WooCommerce, and BigCommerce for commerce; and ServiceTitan, Housecall Pro, and Jobber for field service. The workflow automation tab is built on n8n, self-hosted on the customer's domain.

## Identity and Public Presence

No LinkedIn profile, Crunchbase entry, press coverage, or founder page for this Nikita Rogers has been indexed publicly. The only public signal connecting a founder identity to apples.live is a GitHub account (nikitarogers333) with repos for an AI agent system, a tmux dashboard, and a forked University of San Diego data science and cloud computing course -- suggesting a current or recent USD student with a data science background. A shared SMS contact (917-319-8830) links apples.live to a separate live product, youtubetranscript.us, a YouTube transcript API offered at $9, $29, and $79 monthly tiers. A domain named askcraig.org, registered April 10, 2026 via Namecheap on Cloudflare DNS, is also in the portfolio but its content was not publicly resolvable at the time of this audit.

Social searches for "Nikita Rogers" and related names returned no matching results for this project. The nikitarogers.com domain and the Goodreads author profile belong to a different person entirely: an Urban Fantasy author from Northern Ireland who writes the Whisper of Witches series and works as a tattoo artist. LinkedIn searches returned an unrelated Norwegian computer games company, an Icelandic outerwear brand, and a Berlin-based digital media distribution firm -- none connected to apples.live.

There are no named customers, no case studies, and no testimonials publicly visible on either apples.live or red.apples.name.

Operator + builder. Builds his own platforms (apples, oranges, trading bot). Engineer mindset (sees everything as a system). Family-rooted. Currently focused on scaling AI consultancy + getting first Airbnb foothold + credit recovery. Likes: technical depth, automation, leverage, small experiments that compound. Dislikes: marketing fluff, corporate BS, dead UIs.

Self-content for oranges = reflection / progress mirror / 'what did you actually do this week' rather than entertainment. Could surface: own writing across MD files, own commits, own AI conversations, own metrics from apples/trading. Oranges-for-self is the operator's living dashboard.

{"subreddits": ["realestateinvesting", "airbnb", "personalfinance", "financialindependence", "creditrepair", "Entrepreneur", "StudentLoans", "LocalLLaMA", "sidehustle", "investing", "digitalnomad", "povertyfinance", "automation", "legaladvice"], "youtube_channels": ["Graham Stephan", "BiggerPockets", "Meet Kevin", "Andrei Jikh", "Lex Fridman", "Two Cents", "Andrej Karpathy"], "rss_feeds": ["https://hnrss.org/frontpage", "https://www.inman.com/feed/", "https://techcrunch.com/feed/", "https://rss.nytimes.com/services/xml/rss/nyt/RealEstate.xml", "https://www.entrepreneur.com/latest.rss", "https://www.theverge.com/rss/index.xml", "https://hnrss.org/newest?q=LLM+agents", "https://www.nerdwallet.com/blog/feed/"], "hn_keywords": ["LLM agents", "real estate automation", "credit score", "financial independence", "Airbnb arbitrage", "passive income", "short-term rental", "AI automation", "bootstrapping", "student loans"], "wikipedia_topics": ["Short-term rental", "Credit score in the United States", "Financial independence", "Real estate investing", "Federal Student Aid"], "humor_tone": "dry", "content_length_pref": "mixed", "language": "en", "novelty_appetite": "medium", "hard_avoid": ["multi-level marketing", "crypto speculation", "get-rich-quick schemes", "politics", "family conflict advice", "content requiring large upfront capital", "co-living or house hacking"], "summary": "Systems-thinking Miami entrepreneur bootstrapping short-term rental real estate on FAFSA capital while building AI automation infrastructure, deep into credit repair, financial independence, and leveraging LLMs as daily tools.", "youtube_channel_ids": ["UCa-ckhlKL98F8YXKQ-BALiw", "UCxDXuLvtjJ9Est8PrG3i2VA", "UCPk6nJE8pBZPLCW2DAvaZOA", "UCGy7SkBjcIAgTiwkXEtPnYg", "UCJIfeSCssxSC_Dhc5s7woww", "UChxc_Lg9sh3aYBjVS9h_SHA", "UCYO_jab_esuFRV4b17AJtAw"]}