Getting Started

One path from a checkout to a useful search result.

1. Install dependencies

From the repository root:

$ uv sync --all-groups

2. Search local agent prompts

Ranked search across fast prompt-history stores — deduped, relevance first, with newest as the stable tie-break:

$ uv run agentgrep search "release notes"

Use matching prompts to select and search a bounded set of conversations:

$ uv run agentgrep search "release notes" --deep

Include prompt records stored only inside every readable transcript:

$ uv run agentgrep search "release notes" --exhaustive

Search prompts and conversations together in one explicit deep sweep:

$ uv run agentgrep search "release notes" --exhaustive --scope all

Prefer ripgrep-shaped flags? grep searches prompt-scope records across supported stores:

$ uv run agentgrep grep "release notes"

Search one agent’s prompt records:

$ uv run agentgrep grep "deploy docs" --agent codex

Search full conversation records explicitly. A conversation scope authorizes the transcript reads it requires:

$ uv run agentgrep grep "deploy docs" --agent codex --scope conversations

3. Inspect the stores

See which files and databases agentgrep can read:

$ uv run agentgrep find

Filter discovery output:

$ uv run agentgrep find sessions --agent codex

4. Use MCP

Run the local stdio server:

$ uv run agentgrep-mcp

Or run the FastMCP config:

$ uv run fastmcp run fastmcp.json

See MCP Clients for MCP client snippets.

Next steps

  • Tutorial walks through CLI search in more detail.

  • Tools documents the MCP tool payloads.

  • Configuration explains output, progress, privacy, and source selection.