# Anchor Growth Playbook

Anchor's growth goal is qualified developer interest, not hidden tracking. Keep the public message simple:

```text
Local repo and org memory for AI coding agents.
```

Expanded copy:

```text
Anchor gives AI coding agents local repo and org memory from GitHub PR history, code, tests, regressions, architecture, and cross-repo impact.
```

## MCP Directory Listing Copy

Use this description when submitting or refreshing listings:

```text
Anchor is a local-first MCP server that gives AI coding agents repo and org memory from merged GitHub PR history, current code, tests, regressions, architecture, and cross-repo impact.
```

Recommended categories and tags:

```text
Development
Code Review
Memory
GitHub
Local-first
MCP server
AI coding agents
Repo memory
Org memory
Cross-repo impact
```

Submit or refresh Anchor in:

- Official MCP Registry: https://registry.modelcontextprotocol.io/
- mcpservers.org
- LobeHub MCP
- Smithery
- Glama
- PulseMCP
- awesome-mcp-servers lists
- Relevant GitHub topic pages where submission is possible

Use `docs/directory-submission-pack.md` for ready-to-copy descriptions, tags, links, preview image, and community post drafts.

## Four-Week Campaign

Week 1:

- Merge the repo and org memory positioning update.
- Apply the GitHub About description, website, topics, social preview, and Discussions categories from `docs/github-repo-settings.md`.
- Submit the sitemap in Google Search Console.
- Submit or refresh MCP directory listings.
- Refresh social preview cards after the new `social-preview-repo-org.png` image is live.
- Post a LinkedIn launch update around repo and org memory for AI coding agents.

Week 2:

- Publish a short demo video using `npx @pratik7368patil/anchor demo`.
- Post “Show HN: Anchor, local repo and org memory for AI coding agents.”
- Share practical setup examples in MCP, Cursor, Claude Code, Codex, and developer-tool communities.

Week 3:

- Publish “Why PR history is the missing memory layer for AI coding agents.”
- Share an org-memory post: “What breaks when agents only understand one repo?”

Week 4:

- Publish “Anchor vs code search vs code graphs vs memory MCPs.”
- Ask early users for setup issues, screenshots, use cases, and missing MCP clients.
- Share an aggregate traction update with clear labels and no exact-user claims.

## Metrics To Track

Baseline before rollout:

- GitHub stars, forks, issues, discussions, unique visitors, and unique cloners.
- npm last-day, last-week, last-month, and since-start downloads.
- GoatCounter docs visits and CTA events.
- Google Search Console impressions and clicks.
- MCP directory listing count and referral traffic.
- External mentions from Reddit, Hacker News, LinkedIn, newsletters, and MCP directories.

30-day targets:

- 10+ MCP directory or list placements.
- 25%+ increase in GitHub unique visitors or unique cloners.
- 10+ net new GitHub stars from public traffic.
- 3+ external issues or discussions from non-team users.
- 5+ indexed Google pages for target long-tail queries.
- Search Console impressions trending up for “repo memory MCP,” “GitHub PR history MCP,” and “AI coding agent memory.”

## Measurement Language

Use careful wording:

- “npm downloads are directional demand and include CI, bots, mirrors, and repeat installs.”
- “GitHub unique cloners and visitors are aggregate platform signals, not exact active users.”
- “Anchor does not track CLI usage, local installs, indexed repos, or MCP calls.”
- “Adoption stats are public aggregate signals only.”

## Launch Post Seeds

LinkedIn:

```text
I built Anchor: local repo and org memory for AI coding agents.

It indexes GitHub PR history, current code, tests, regressions, architecture, and cross-repo impact locally, then exposes concise cited context through MCP.

No SaaS. No CLI telemetry. Read-only GitHub access. SQLite on your machine.

Try the offline demo:
npx @pratik7368patil/anchor demo
```

Show HN:

```text
Show HN: Anchor, local repo and org memory for AI coding agents

Anchor is a local-first MCP server and CLI that indexes merged GitHub PR history plus current code, tests, regressions, architecture, and explicitly allowlisted org repos. It helps coding agents retrieve evidence before edits and check cross-repo impact before API/shared-package changes.

It has no SaaS, no CLI telemetry, no GitHub write access, and no remote LLM calls.

Demo:
npx @pratik7368patil/anchor demo
```