Record and query debug sessions, commands, and fixes in a local SQLite database. Built for developers who want to persist and search their troubleshooting history across projects.
io.github.oaslananka/mcp-debug-recorder
Local install
STDIO
No auth required
How models use it and what it is built for.
Record and query debug sessions, commands, and fixes in a local SQLite database. Built for developers who want to persist and search their troubleshooting history across projects.
Local install — runs as a subprocess.
Configuration this server reads at startup.
Override the SQLite database location for local debug history.
Set the minimum structured log level: debug, info, warn, or error.
Override the Fuse.js threshold used during fuzzy reranking.
Where to find authoritative docs and source for MCP Debug Recorder.
Paste any of these into Agent Studio after connecting MCP Debug Recorder.
Common questions about connecting and running MCP Debug Recorder.
Where does the debug recorder store my session data?
By default, it stores debug history in a local SQLite database. You can override the database location by setting the DEBUG_RECORDER_DB environment variable to a custom path.
How do I install and run the MCP debug recorder?
Install it with `npx mcp-debug-recorder@1.0.1`. It runs over stdio transport and requires no external API keys—all data stays local in your SQLite database.
Can I control how verbose the logs are?
Yes. Set the LOG_LEVEL environment variable to debug, info, warn, or error to control the minimum structured log level recorded.
What does the FUZZY_THRESHOLD setting do?
It controls the sensitivity of fuzzy search when reranking results in your debug history. Override it via the FUZZY_THRESHOLD environment variable to tune how closely search terms must match recorded sessions.
Is my debug history private and local-only?
Yes. All debug sessions, commands, and fixes are stored in a local SQLite database on your machine—nothing is sent to external servers.
MCP Playground runs 10,000+ hosted MCP servers — GitHub, Linear, Notion, Stripe, Sentry and more — across Claude, GPT, Gemini, DeepSeek and 30+ AI models. Compare model answers side-by-side, save agent presets, share runs. Zero install.
Open Agent Studio