{"$schema":"https://wellknown.network/schemas/agent-record-v1.json","schemaVersion":"1","id":"ag_zvsaee5rd83w","handle":"librarian-mcp","url":"https://wellknown.network/agents/librarian-mcp","links":{"self":"https://wellknown.network/agents/librarian-mcp/record.json","html":"https://wellknown.network/agents/librarian-mcp","markdown":"https://wellknown.network/agents/librarian-mcp/record.md","api":"https://wellknown.network/api/v1/agents/librarian-mcp","status":"https://wellknown.network/api/v1/agents/librarian-mcp/status","claim":"https://wellknown.network/agents/librarian-mcp/claim","claimApi":"https://wellknown.network/api/v1/claims","claimDescriptor":"https://wellknown.network/agents/librarian-mcp/claim.json","badge":"https://wellknown.network/agents/librarian-mcp/badge.svg","openapi":"https://wellknown.network/openapi.json","history":"https://wellknown.network/api/v1/agents/librarian-mcp/history","tools":"https://wellknown.network/api/v1/agents/librarian-mcp/tools"},"ard":{"identifier":"urn:air::server:librarian-mcp","type":"application/mcp-server-card+json"},"kind":"mcp_server","declared":{"name":"librarian-mcp","summary":"Pre-curated canonical memory + prose/code provenance for any AI coding client. MCP-native.","description":"# Librarian MCP\n\n**A real, measured alternative to \"bigger context windows.\"**\nPre-curated canonical memory + prose/code provenance checking + benchmark metrics, delivered as a [Model Context Protocol](https://modelcontextprotocol.io) server that works across Claude Code, Cursor, VSCode (via Continue), and any MCP-capable client.\n\n## What it does\n\nFive tools, all exposed via MCP:\n\n| Tool | What it does | Added |\n|---|---|---|\n| **`librarian_context`** | Intent-aware canonical memory packet. Loads curated preload content scoped to your query intent (outreach, architecture, benchmark, founder voice, etc.). Eliminates the \"forgets by prompt #21\" failure mode. | v0.1.0 (stub), **v0.2.0** (intent-aware) |\n| **`prose_provenance`** | Deterministic drift detection between two document versions. Catches silently-removed voice anchors, stale canonical numbers, section changes, register shifts. | v0.1.0 |\n| **`record_measurement`** | Log a single benchmark measurement (vendor, model, condition, accuracy, cost, latency) to local JSONL. | **v0.2.0** |\n| **`metrics_summary`** | Per-vendor and per-model aggregation of recorded measurements. Shows accuracy lift, cost savings, cache hit rate. | **v0.2.0** |\n| **`opt_in_share`** | Toggle anonymous metrics sharing flag. Default OFF. Commons dashboard POST endpoint ships in a future release. | **v0.2.0** |\n\n## Why we built this\n\nIndependently measured result (Eyewitness Benchmark R10, April 2026, eight models across four vendors, 1,200 graded calls, inter-rater kappa 0.883/0.850):\n\n- **Without the Librarian (COLD):** mean 8.7% correct\n- **With the Librarian (HOT):** mean 94.8% correct — **86.1 percentage-point lift**\n- **Haiku 4.5 (cheapest) ties Opus 4.7 (most expensive)** at 19x cost difference\n- **4.3x more right answers per dollar of compute**\n\nApplied inside Microsoft Copilot's inference path, the same architecture recovers an estimated **$750M/year** in waste. Inside Anthropic's developer tools, **~$130M/year**. Full methodology…","publisher":null,"homepage":"https://github.com/liana-banyan/librarian-mcp/tree/main/docs","repository":"https://github.com/liana-banyan/librarian-mcp/tree/main/docs","version":"0.2.0","license":null,"protocols":["mcp"],"tags":["ai","benchmark","canonical","context","llm","mcp","memory","provenance"],"pricing":null,"endpoints":[{"url":"pypi:librarian-mcp","type":"package_pypi","auth":null,"probeable":false}],"skills":null,"tools":null,"extra":null,"attribution":{"kind":"pypi","name":"pypi","repoUrl":"pypi","summary":"pypi","version":"pypi","description":"pypi","homepageUrl":"pypi"}},"derived":{"capabilities":[{"slug":"knowledge.memory","name":"Agent Memory","confidence":1,"provenance":"declared"},{"slug":"ai.evaluation","name":"Evaluation & Benchmarks","confidence":1,"provenance":"declared"}],"categories":["ai","knowledge"],"language":"en"},"observed":{"status":"unknown","statusReason":"Distributed as a package to run locally; no network endpoint to check.","lastOkAt":null,"lastProbedAt":null,"statusComputedAt":null,"reliability30d":null,"latestObservations":[],"tools":null,"package":{"name":"librarian-mcp","registry":"pypi","observedAt":"2026-10-28T08:28:18.844Z","publishedAt":"2026-04-22T20:25:56.602698Z","latestVersion":"0.2.0"},"toolSurface":null,"endpointFacts":[]},"verification":{"claimed":false,"claimedAt":null,"proofs":[]},"provenance":{"sources":[{"source":"pypi","key":"librarian-mcp","url":"https://pypi.org/project/librarian-mcp/","firstSeenAt":"2026-09-09T23:25:23.491Z","fetchedAt":"2026-10-28T08:27:47.468Z","normalizedAt":"2026-10-28T08:27:47.468Z"}]},"firstSeenAt":"2026-09-09T23:25:23.491Z","updatedAt":"2026-10-28T08:28:18.844Z"}