Execution · July → September 2026

Ninety days to first revenue and a funnel that compounds.

The 90-day job is not launching a business — it's building the three assets that unlock 2027: a paying client + case study, a live funnel, and a metering skeleton.

01 /

Month 1 — Package & polish

july

Weeks 1–4

Goal: a sellable offer and an install experience that doesn't leak users.

  • Write the productized offer — "orchestration setup + 3 custom workflows, $500–$2K fixed" — and put it on the landing page with a booking link.avenue 1 · the only near-term cash
  • Ship the usability quick wins: doc/env consistency (GEMINI_API_KEY, Node version, default profile), doctor/tachi diagnose wired to ToolRouter.getStatus, catalog mode, verify-installation OPENROUTER check.docs quality = #1 sponsor/Pro conversion driver; broken onboarding kills every avenue
  • Add the Claude Code one-liner install to the README top (claude mcp add … npx).2026 install norm — competitors all meet it
  • Turn on GitHub Sponsors + FUNDING.yml; open a Polar account.avenue 5 · free trust signal
  • Submit to directories: Anthropic Connectors, Smithery, Glama, PulseMCP, mcp.so — freemium posture.avenue 4 · pure distribution
02 /

Month 2 — Funnel & first client

august

Weeks 5–8

Goal: one paying engagement, one Pro-tier gate live.

  • Land 1–2 service clients — direct outreach to teams adopting MCP; the offer is already public.avenue 1 · $1K–5K
  • Build the usage-analytics dashboard (real token capture, per-provider spend) and gate it as the Polar Pro tier.avenue 2 · the credibly-wanted gate; also a product differentiator
  • Publish 2 free pieces on the jury/council protocols and zero-cost local jurors.avenue 6 · top-of-funnel for everything
  • Add a CLA note to CONTRIBUTING to keep sole copyright → dual-license optionality.avenue 9 · preserve the AGPL moat
03 /

Month 3 — Skeleton & proof

september

Weeks 9–13

Goal: the metering loop works end-to-end; the case study is public.

  • Publish the client case study on the landing page.converts the next client at higher price
  • Stand up the hosted-tier skeleton: thin managed deploy, holds keys, meters margin over OpenRouter, billed via the same Polar sub. No public launch — steps and platform pick on the Cloud & Hosting tab.avenue 3 · proves the 2027 margin loop · no AWS needed
  • Wire the budget cap into live calls (cost-monitor exists, unwired) and expose spend in usage_stats.trust feature BYOK users need before they'll ever hand you keys
  • Review the funnel numbers and decide Q4: double down on services vs. push hosted beta.decision gate, not a guess
04 /

Roadmap = product roadmap

same work, two payoffs
Usability featureAdoption payoffMonetization payoff
npx tachibot initCollapses 7-step onboarding to one commandMore installs → bigger funnel for every avenue
doctor / diagnoseSilent missing-key tool absence becomes actionableTrust prerequisite for Pro tier conversion
real cost trackingBYOK spend visibility no competitor hasThe analytics dashboard gate for Polar Pro
budget caps (wire cost-monitor)Safe to use expensive modelsRequired before anyone trusts TachiBot Cloud with keys
conversation continuityCloses the headline gap with Zen MCPRetention → recurring subscription value
05 /

Metrics that matter

check monthly
1st
paid service engagement — the quarter's defining metric
npm ↑
weekly downloads trend (already on the landing badge)
★ 15→?
GitHub stars after directory listings + content pieces
≥1
Polar Pro subscriber proves the gate; 6% of free users is the benchmark