MCP.so
Sign In

Selvedge MCP Server

@masondelan

Basic information

About Selvedge MCP Server

Decision provenance for AI-coded codebases: the why, and what was already tried and rejected. Exports and imports Agent Trace v0.1.0, so captured history round-trips with the emerging open standard for AI code attributi

Selvedge MCP Server config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "selvedge": {
      "command": "uvx",
      "args": [
        "--from",
        "selvedge",
        "selvedge-server"
      ]
    }
  }
}

Selvedge MCP Server tools

Tool list not available yet

This listing does not currently have a tool list to display. This does not mean the server has no tools; check its documentation for supported capabilities.

Fetch the live tool list by running this server in a temporary sandbox using the button above.

Overview

What is Selvedge?

Persistent decision memory for AI coding agents. Save why you chose an approach, what you rejected, and what would change your mind. Retrieve that context in the next coding session, using a function, database column, API route, or dependency as the lookup key.

Selvedge is a local MCP server and CLI. Claude Code, Codex, Cursor, Copilot, Gemini CLI, and Windsurf can connect to the same SQLite store in .selvedge/. No hosted account is required; core storage and retrieval make no LLM calls. Only explicitly saved decisions can be recalled. Recorded explanations are supplied by a person or agent and are not verified facts.

Get started

Requires Python 3.10+ and uv.

uv tool install --upgrade selvedge
selvedge demo
cd your-project
selvedge setup --agent codex

Use codex, claude-code, cursor, copilot, gemini, or windsurf. The demo uses a temporary database. After setup, restart your coding client, save a real decision, and retrieve it in a new session. See the first-decision verification guide and agent compatibility reference.

Current release: v0.3.16

  • Eight MCP tools: log_change, prior_attempts, blame, diff, history, changeset, search, and stale_decisions.
  • Entity-level decision history, rejected approaches, append-only superseding records, changesets, and revisit conditions.
  • selvedge ledger shows recorded actor/session attribution and explicit revisions. Attribution is self-reported.
  • selvedge doctor --agent CLIENT checks project hook configuration for all six setup targets. Configuration checks do not prove that a client activated its hooks.
  • Optional PR review Action carries approved decision history into reviews without executing PR-head code.
  • Native lifecycle adapters cover supported client events; startup context, watched-edit checks, and compaction notifications vary by harness.
  • Agent Trace v0.1.0 import/export and Markdown, JSON, CSV exports.

Release and source · Documentation · PyPI

How does it differ from git history?

Git records code changes. Selvedge stores the stated reasons, alternatives, and conditions behind decisions, then makes them queryable before later edits. It does not guarantee that an agent will retrieve or correctly apply them.

Where is data stored?

In .selvedge/selvedge.db in the project, with a global fallback under ~/.selvedge/. MCP transport is JSON-RPC over stdio. Optional integrations and semantic search are separate from the local core.

Comments

More AI & Agents MCP servers