Skip to main content
Glama
azerv1
by azerv1

thimiko

thimiko is θυμικό.

It makes your local AI coding chats searchable. It reads Codex, Claude Code, GitHub Copilot, and Gemini CLI history, builds a local SQLite index, and lets you search it from the terminal or through MCP.

Your original chat files are left alone.

Why

Useful things get buried in old chats: fixes, decisions, failed approaches, commands, and the one detail you need six weeks later.

With thimiko, you can ask things like:

  • "Have I leaked a secret that starts with 0xSecret in our chats?"

  • "Did we talk about X last week?"

  • "Can you collect everything from our chats about Y and make me a study plan?"

  • "Why did we decide not to use Redis for the job queue?"

  • "Find the chat where we fixed the flaky migration test."

  • "What command did we use to recover the broken production index?"

  • "Have I pasted any customer names, internal URLs, or production hostnames?"

  • "Show me previous attempts at this error, including the surrounding messages."

It is especially handy for secret audits, incident reconstruction, architecture decision archaeology, compliance checks, recurring bugs, and picking up an old project without starting from zero.

Related MCP server: RepoChatMCP

Install

uv tool install "thimiko[mcp]"

For local development:

uv tool install -e ".[mcp]"

Use it

Run thimiko by itself to see the available commands.

Build the index once:

thimiko build

Then keep it fresh and search it:

thimiko update
thimiko search "database migration"
thimiko search "permission denied" --source codex --limit 5
thimiko search "rate limit" --days 10 --text

By default, thimiko reads:

  • Codex: ~/.codex/sessions/**/*.jsonl

  • Claude Code: ~/.claude/projects/**/*.jsonl

  • GitHub Copilot: VS Code's workspaceStorage/*/chatSessions/

  • Gemini CLI: ~/.gemini/tmp/*/chats/

You can also pass a file or directory directly to build or update.

MCP

Run the read-only MCP server with:

thimiko mcp

It exposes three tools: search_chats, get_turn, and get_session. Point your MCP client at thimiko mcp, then your assistant can search old chats and open the relevant surrounding conversation.

The index stays local. Its default location is %LOCALAPPDATA%\thimiko\thimiko.sqlite on Windows.

For implementation details and instructions for adding another chat source, see ARCHITECTURE.md.

A
license - permissive license
-
quality - not tested
B
maintenance

Maintenance

Maintainers
Response time
Release cycle
Releases (12mo)
Commit activity

Resources

Unclaimed servers have limited discoverability.

Looking for Admin?

If you are the server author, to access and configure the admin panel.

Related MCP Servers

  • A
    license
    A
    quality
    D
    maintenance
    Enables searching and analyzing GitHub Copilot's conversation history stored locally, providing tools for full-text search, session listing, statistics, and file-based retrieval.
    Last updated
    6
    4
    MIT
  • A
    license
    -
    quality
    D
    maintenance
    A zero-dependency MCP server that enables searching and reading local Claude Code and Codex chat sessions, supporting full-text search, grep, and knowledge indexing from chat history.
    Last updated
    5
    MIT
  • F
    license
    -
    quality
    F
    maintenance
    Full-text search over Claude Code conversation history using SQLite FTS5, exposing indexed transcripts as MCP tools for searching, browsing, and reading turns.
    Last updated
    3

View all related MCP servers

Related MCP Connectors

  • Search your AI chat history (ChatGPT, Claude, Codex) from any MCP client. Remote, private, read-only

  • Agentic search over your Dewey document collections from any MCP-compatible client.

  • Local-first RAG engine with MCP server for AI agent integration.

View all MCP Connectors

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/azerv1/thimiko'

If you have feedback or need assistance with the MCP directory API, please join our Discord server