Claude-tinderbox: Search your Claude.ai conversation history locally via MCP

A developer created Tinderbox, a local search system for Claude.ai conversation archives that allows any Claude session to query past conversations via two MCP tools. The system ingests exported conversation ZIP files into a Supabase database with hybrid semantic and full-text retrieval, achieving 68.7% top-1 accuracy on a 150-query test set across 676 conversations and 10,653 messages.

A personal claude.ai conversation archive — schema, ingest, embeddings, hybrid retrieval, and an MCP server that lets any Claude session search your own past conversations. Status:working end-to-end. Used daily by the author. Not packaged for general consumption — seeCaveatsbelow. You export your conversations from claude.ai, drop the ZIP into a watched directory, and within ~15 minutes your full archive is searchable from any Claude session via two MCP tools: tinderbox search query, limit=10 — hybrid semantic + full-text retrieval over every message and artifact tinderbox get conversation export id, max messages=50 — pull the full thread of any conversation surfaced by a search Everything is local-ish — you bring your own Supabase free-tier project, your own Ollama install for embeddings, and a Mac this is targeted at Apple Silicon . The current author's archive: 676 conversations, 10,653 messages, 172 artifacts, 10,731 mxbai-embed-large vectors. Hybrid retrieval hits 68.7% top-1 / 88.7% top-10 on a frozen 150-query QA set generated by Haiku from the corpus itself re-runs weekly via launchd . Tinderbox stores statements, not facts. Every retrieval response renders the provenance inline — never "X is true" , always "on date , in conversation , participant said content " . The corpus answerswhat was said when, by whom; neverwhat is true. That's design principle 1 from the schema doc /luckyrmp/tinderbox-archive/blob/main/docs/STAGE 1 SCHEMA PROPOSAL.md . Memorial archive, not extraction pipeline. Forward-linked when superseded, never backward-edited. For context-window reasons it's also genuinely useful: a Claude session can look up its own past reasoning instead of re-deriving it. Postgres Supabase holds 12 tables under a tinderbox schema — schema versioning, ingest runs, conversations, messages, artifacts, attachments, embeddings vector 1024 + hnsw , enrichment, named instances, query log, and a frozen QA test set. A Python parser stream-reads claude.ai export ZIPs and upserts everything idempotently. An embed worker batches messages and artifacts through Ollama mxbai-embed-large , 1024-dim and writes vectors back. A server-side Postgres function tinderbox.hybrid search ranks results by 1 - cosine distance + 0.5 ts rank cd . A small from-scratch JSON-RPC 2.0 MCP server exposes two tools over stdio. Three launchd daemons run the whole thing on a schedule: inbox watcher 15min , QA eval Sundays 03:00 , staleness alerter daily 09:00 with cooldown + debounce . macOS with Apple Silicon, Python 3.14 or 3.12+ probably — author runs 3.14.3 1 Supabase free-tier project — $0/month for this scale. Optional $4/mo IPv4 add-on https://supabase.com/pricing for proper RLS scoping stage 5b . Ollama running locally with mxbai-embed-large pulled ollama pull mxbai-embed-large - A claude.ai data export ZIP Settings → Account → Export Data 1. Clone git clone <this repo ~/tinderbox && cd ~/tinderbox 2. Create your Supabase project, get URL + service-role key 3. Render config + plists for your $HOME / $USER ./parser/scripts/setup.sh 4. Create your env file path is configurable via TINDERBOX ENV FILE cp .env.example ~/.tinderbox.env … and fill in SUPABASE URL, SUPABASE SERVICE KEY, etc. 5. Apply the migrations to your Supabase project each migration file is plain SQL — run them in order via the Supabase SQL editor, or via psql, or via your tool of choice ls migrations/ 6. Pull the embedding model ollama pull mxbai-embed-large 7. Set up the venv the project uses a .pth bridge to share deps from other venvs on the author's machine; you'll likely want to install fresh — see parser/pyproject.toml for the deps list python3 -m venv parser/venv parser/venv/bin/pip install supabase python-dotenv click httpx pydantic anthropic 8. Drop your export ZIP into the inbox and watch it ingest mkdir -p inbox mv ~/Downloads/data- .zip inbox/ parser/venv/bin/python -m tinderbox.cli scan-inbox 9. Embed everything parser/venv/bin/python -m tinderbox.cli embed 10. Try a search parser/venv/bin/python -m tinderbox.cli search "your test query" 11. Wire to Claude Code / Desktop — see docs/MCP INSTALL.md Activate the daemons optional but recommended : launchctl load ~/Library/LaunchAgents/com.$USER.tinderbox.scan.plist launchctl load ~/Library/LaunchAgents/com.$USER.tinderbox.qa.plist launchctl load ~/Library/LaunchAgents/com.$USER.tinderbox.staleness.plist . ├── migrations/ Numbered SQL — apply to your Supabase project in order ├── parser/ │ ├── tinderbox/ Python package │ │ ├── parser/ ZIP → typed records streaming JSON, content-block parsing, artifact versioning │ │ ├── ingest/ Records → DB upsert, retry, tombstone sweep, mass-tombstone canary │ │ ├── embed/ mxbai-embed-large via Ollama, batched, idempotent, per-row fallback │ │ ├── search/ Hybrid retrieval + query logging │ │ ├── qa/ Frozen-query-set eval Haiku-generated, scheduled │ │ ├── mcp/ Minimal JSON-RPC 2.0 MCP server no SDK dep │ │ ├── staleness.py Daily check w/ cooldown + debounce │ │ ├── cli.py tinderbox <command │ │ └── ... │ ├── tests/ │ ├── scripts/ Setup, MCP launcher, surgical recovery │ └── launchd/templates/ Plist templates filled by setup.sh ├── docs/ │ ├── STAGE 1 SCHEMA PROPOSAL.md Design principles + table-by-table rationale │ ├── STAGE 1 COMPLETION REPORT.md Bugs found + fixed during ingest │ ├── STAGE 2 COMPLETION REPORT.md Embed + hybrid retrieval shipped │ ├── STAGE 5 COMPLETION REPORT.md MCP server + IPv4 add-on │ ├── ACCEPTED ADVISORIES.md Supabase advisor findings + accepted/applied/deferred │ ├── MCP INSTALL.md Claude Code / Desktop config snippets │ └── STAGE 2 HANDOFF.md Inter-session handback historical └── README.md this file Hardcoded paths. The author runs everything under ~/tinderbox/ . parser/scripts/setup.sh renders the launchd plists for your $HOME / $USER but the Python defaults still assume that root. TINDERBOX env vars override every default — set them up in your env file. No tests for end-to-end MCP from a real client. The smoke test in parser/tests/test mcp smoke.py spawns the server as a subprocess and exchanges minimal protocol. Real validation is "does Claude Code surface the tool" verified and "does the tool return useful results" eyeballed .The MCP server still authenticates via service role bypass for stage-1. service role RLS bypassed . Documented in docs/STAGE 5 COMPLETION REPORT.md — fine until you start differentiating privacy classes; then stage 5b auth swap to tinderbox owner direct connection becomes urgent. No SQLite option. Supabase only. The free tier easily handles this scale; if you want 100% local, you'll need to translate the schema and rewrite the DB layer ~5-6 hrs of work — author chose not to . macOS only. launchd schedules, paths, and the .pth venv bridge are macOS conventions. Linux would need systemd units and a different venv approach. Not difficult, just not done. Author's archive shape baked into a few decisions. The 5 MB RAW CONTENT BYTE LIMIT was chosen because the author's largest message is 57 MB. The stratified sampler in qa/sample.py uses bucket sizes long ≥20 msgs, short 3-10 msgs tuned to the author's distribution. Both are easy to retune. Pick one. The author is open to anything that lets people fork and adapt without obligation. Built collaboratively across many Claude sessions over a few days in April 2026. Schema design → parser → embed → search → QA → MCP server, mostly autonomous, with the human Lucky stepping in at architecture decisions and pushing back when something didn't smell right. The system can now query the very conversations that built it.