MCP server that indexes a code repository into a content-addressed graph, then returns ranked code context for tasks such as impact analysis, test selection, and symbol lookup. It also supports cryptographic proofs and offline verification of code relationships from graph snapshots tied to git commits.
From the registry: Content-addressed code graph with 22 MCP tools for AI agents.

$ brew install blackwell-systems/tap/knowingPlease install the `knowing` MCP server into my current AI client (that's you).
Required prerequisites (do these first if not already done):
- **Install knowing** — Install the knowing CLI/MCP server. The README presents Homebrew as the primary install method and also lists npm, pip, and go install alternatives. Run: `brew install blackwell-systems/tap/knowing`
Canonical MCP server config (stdio transport):
- command: `knowing`
- args: ["mcp","--watch"]
Note: Preferred MCP setup uses stdio with `knowing mcp --watch`. The server auto-indexes the current git repository on first launch. It also auto-downloads a local embedding model (~30MB) on first use; no API keys are required. The README also documents an HTTP mode via `knowing serve -addr :8100 .`, but stdio is the recommended/default path. To disable embeddings, add `--no-embeddings` to the stdio args.
Add this MCP server to my current client's config in the correct format for you. If you need secrets or credentials I haven't provided, ASK me — do not invent values or leave raw placeholders. After adding it, tell me how to verify the server is connected.Connect AI assistants to GitHub - manage repos, issues, PRs, and workflows through natural language.
The GitKraken MCP Server for managing repos, PRs, issues across GitHub, GitLab, Bitbucket and more.
Comprehensive Git MCP server enabling native git tools including clone, commit, worktree, & more.