Concord AI
Concord AI builds shared work-state and task-memory infrastructure for coding agents. Its stated goal is to let Claude Code, Codex, Cursor and other MCP-capable agents coordinate presence, claims, messages, decisions, ownership, handoffs and review evidence before code reaches pull-request review.
At a Glance
- Software-development teams using AI coding agents
- Engineering teams running multiple agents or multiple coding harnesses
- Developers using Claude Code, Codex, Cursor, Gemini CLI, Grok Build or other MCP-capable clients
- Teams that want repository-local, inspectable coordination state
AI Tools by Concord AI
(1)
Concord MCP
MCP Server for Multi Agent Coding
Discussions
No discussions yet
Be the first to start a discussion about Concord AI
Latest News
How an AI native engineering team actually works
How to pair program with an AI coding agent
Cursor rules that actually work for coding agents
Concord MCP v0.10.3 released with persisted task outcomes and outcome-driven telemetry v2
Products & Services
An open-source, local-first MCP server and CLI that provides a shared SQLite work-state inside a repository. It enables presence, pre-edit scope claims and overlap detection, cross-harness messaging, task memory, explicit ownership transfers, structured handoffs and review packets. Distributed as the MIT-licensed npm package @concord-ai/concord-mcp.
A planned service described on the official website as coming soon and intended to coordinate agents across machines, in contrast with Concord MCP's repository-local OSS mode.
Market Position
Concord positions itself as the shared coordination layer around coding agents rather than another autonomous agent or orchestration system. Its differentiators are local-first repository state, pre-edit overlap detection, durable task memory and evidence-based handoffs; the alternatives it implicitly addresses are unstructured chat, shared files, Git branches/worktrees and late pull-request reconciliation, while its planned Cloud offering would extend the model across machines.
Leadership
Founders
Alex Choi
Founding engineer associated with Concord AI; GitHub describes him as a full-stack and AI developer based in London, and his profile says he works across AI/ML, frontend, backend and mobile development.
Executive Team
Alex Choi
Founding Engineer / AI Engineer
Builds tooling for teams whose code is mostly written by coding agents; GitHub identifies him as a full-stack and AI developer and says he is currently working on Concord AI.
Albin Jaldevik
AI Engineer
Works on agent workflows, review evidence and keeping generated code reviewable. His personal site describes him as a quantitative trader at DRW with an M.Sc. in Computer Science (AI) and degrees in mathematics and engineering.
Business Model
Revenue Model
The currently available product is an MIT-licensed open-source, local-first MCP server distributed through npm. The official site presents Concord Cloud, a cross-machine hosted offering, as coming soon; no public pricing or revenue figures were found.
Target Markets
- Software-development teams using AI coding agents
- Engineering teams running multiple agents or multiple coding harnesses
- Developers using Claude Code, Codex, Cursor, Gemini CLI, Grok Build or other MCP-capable clients
- Teams that want repository-local, inspectable coordination state
- Coordinating multiple AI coding agents working in parallel on one repository
- Preventing overlapping edits and late merge conflicts
- Moving unfinished tasks between agents or coding harnesses without losing context
- Preserving decisions and assumptions across sessions
- Preparing AI-generated code for human review with structured evidence
- Running local-first agent collaboration without a hosted service