ContextPact
Allows Obsidian to edit the human-readable Markdown layer of a ContextPact workspace, with planned reconciliation, import/export, backup and recovery.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@ContextPactShow approved decisions and task handoffs from my ContextPact workspace."
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
ContextPact
Any agent. Same approved context.
ContextPact is a local-first context and coordination layer for people who move between AI tools or run several agents in parallel. It is designed to keep durable knowledge, current decisions, task ownership and handoffs outside any one AI product.
Project status
ContextPact is under active pre-release development. The repository currently contains the approved v1 specification, architecture, repository foundation and an experimental workspace bootstrap. It is not yet a production-ready context system and has not been released to npm.
Implemented and tested in the current foundation:
TypeScript build and test pipeline.
Context item and workspace contracts.
Local workspace layout creation.
SQLite schema bootstrap with WAL and FTS5.
Experimental
init,statusand MCP bootstrap/status surfaces.
Planned for v1.0 and not yet advertised as supported:
Complete knowledge lifecycle, context-pack retrieval and supersession.
Task leases, conflict handling and structured handoffs.
Guided Claude Code, Codex and Cursor connection.
Obsidian edit reconciliation, import/export, backup and recovery.
Cross-platform clean-install and end-to-end proof.
Related MCP server: MCP Project Bridge
Intended experience
contextpact init
contextpact connect claude
contextpact connect codex
contextpact statusThe connection commands above describe the planned v1 experience; only the documented foundation commands should be treated as implemented before the release checklist passes.
Architecture
Markdown owns human-readable knowledge, rules, goals, decisions and handoff narratives.
SQLite owns agents, sessions, tasks, leases, policies, versions and audit events.
FTS5 is a rebuildable search index; semantic retrieval is optional future work.
CLI and MCP call the same transport-independent core.
Obsidian may edit the Markdown layer but is never required.
See the product specification, architecture, roadmap and release checklist.
Development
Requires Node.js 22.12 or newer.
npm install
npm run check
npm run dev -- init ./scratch-workspace
npm run dev -- status ./scratch-workspaceLicence
MIT. See LICENSE.
This server cannot be deployed
Maintenance
Related MCP Connectors
Markdown workspace for AI agents: read, write, organize, and share markdown documents.
Persistent docs and memory for AI agents — read, write, organize & search a shared workspace.
Durable agent-to-agent handoffs and shared scratchpad for multi-agent workflows.
Cross-agent artifact workspace with provenance across Claude Code, Codex, Cursor, LangGraph.
Related MCP Servers
- AlicenseNot gradedqualityAmaintenanceLocal-first, file-based memory layer for AI agents — one shared Markdown vault across Claude, Codex, Gemini, Cursor and any MCP client. Provides read/write memory tools with an audit trail, per-agent trust levels, and Git sync; no cloud and no lock-in.2MIT
- AlicenseNot gradedqualityCmaintenanceEnables multi-project workspaces to share structured notes, API contracts, and handoff messages via a local SQLite database, with versioning and read tracking.GPL 3.0
- AlicenseNot gradedqualityCmaintenanceEnables AI agents to read and write a local-first knowledge base of plain markdown files in git, with governance gates for safe, hash-anchored edits.1Apache 2.0
- AlicenseAqualityBmaintenanceA local-first shared memory layer for MCP-aware agents like Claude, Codex, and Hermes, enabling persistent memory across chats and clients via Markdown files and SQLite FTS.62MIT