Rutter

Rutter records AI session decisions and cited notes in a local Markdown folder using an MCP server.

Rutter is a tool that stores past AI decisions and the notes they cite in dated Markdown files. As an AI session concludes a decision, your client appends a summary line containing touched paths and file content hashes. It works with local knowledge vaults, such as Obsidian, and functions as an MCP server. It is built for individuals who keep Markdown notes and configure AI clients. Supported AI clients include Claude Code, Grok, Codex, and Antigravity. Rutter sets itself apart by keeping records outside of any single proprietary tool, allowing different AI assistants to access the same chronological history.

Rutter at a glance

Best for: Software engineer
Runs on: Web, API, CLI, self-hosted, open source
Pricing: Free and open source under the MIT license with no support commitment.
Released: October 2026
Category: AI coding tool
Website: ruttermemory.com

Key features

Appends session decisions to dated Markdown records
Stores content hashes of cited notes at capture time
Resolves renamed note paths automatically via file hashes
Operates as an MCP server for multiple AI clients
Stores working directories and git remotes for context
Reindexes positions across different sessions and tools

Use it for

Recall past decisions across different AI tools
Track which notes a previous AI session cited
Maintain a chronological log of knowledge work
Follow note path renames in AI session history

Alternatives to Rutter

AI Memory — Shared memory for coding agents that works across machines and team members.
Rubi — It cuts AI token waste by keeping sessions short and mapping project knowledge.
Juggler — A native Go server with CRDT-based live conversations and recursive branching threads.
Agent Archive — Hand off coding sessions between Claude Code, Codex, and Cursor across machines.
Marmot — Consolidates multiple AI and data provider interfaces into one CLI.
Mdt — Avoids unnecessary rereading, token use, and distraction for agents.
Engram — Single Go binary with zero dependencies and built-in full-text search.
Crit — It keeps human feedback localized and precise for AI agent code iterations.
DocuWriter.ai — Generates and maintains up-to-date documentation directly from code.
Memorax Code — Carry project architecture and verified engineering decisions across coding sessions.
Enzyme — It indexes markdown notes locally so AI coding agents can cite them.
DeployRamp — Automates feature flag wrapping, rollout monitoring, and cleanup for code.

All Rutter alternatives