LaunchChair MCP turns project work into a local agent loop.
The LaunchChair MCP bridge runs locally over stdio. Desktop agents call MCP tools, the bridge talks to LaunchChair Agent API, and a configured runner executes the queued work.
Install
Download the local bridge files.
The bridge files are public JavaScript files. Download them locally, then configure your MCP client with the bridge path and your authenticated LaunchChair environment variables.
mkdir -p "$HOME/launchchair-agent" curl -fsSL "https://www.launchchair.io/api/agent/v1/bridge/mcp-bridge-server.mjs" -o "$HOME/launchchair-agent/mcp-bridge-server.mjs" curl -fsSL "https://www.launchchair.io/api/agent/v1/bridge/runner-command.mjs" -o "$HOME/launchchair-agent/runner-command.mjs" curl -fsSL "https://www.launchchair.io/api/agent/v1/bridge/codex-runner.mjs" -o "$HOME/launchchair-agent/codex-runner.mjs" curl -fsSL "https://www.launchchair.io/api/agent/v1/bridge/claude-code-runner.mjs" -o "$HOME/launchchair-agent/claude-code-runner.mjs" curl -fsSL "https://www.launchchair.io/api/agent/v1/bridge/chatgpt-runner.mjs" -o "$HOME/launchchair-agent/chatgpt-runner.mjs" curl -fsSL "https://www.launchchair.io/api/agent/v1/bridge/claude-runner.mjs" -o "$HOME/launchchair-agent/claude-runner.mjs" chmod +x "$HOME/launchchair-agent"/*.mjs
Environment
The bridge needs a base URL, token, project ID, and runner target.
The public docs use placeholders. Users get the real token and project ID from Settings -> API Access after logging in.
LAUNCHCHAIR_BASE_URL="https://www.launchchair.io" LAUNCHCHAIR_AGENT_TOKEN="<full lc_at token from Settings API Access>" LAUNCHCHAIR_PROJECT_ID="PASTE_PROJECT_UUID"
Use LC_AGENT_ROUTER_GPT_TARGET and LC_AGENT_ROUTER_CODEX_TARGET to route work.
Use codex for Codex Desktop local runs.
Use claude_code for Claude Desktop plus Claude Code local runs.
Set the CWD variable to the repo LaunchChair should build into.
Tools
The primary loop tool is launchchair_continue_project.
The loop tool asks LaunchChair to inspect project state, queue the next agent-actionable step, run it with the configured local runner, complete it, apply or validate output, remediate when needed, and return phase synthesis links when phases finish.
Call launchchair_continue_project with:
{
"projectId": "PASTE_PROJECT_UUID",
"maxSteps": 8,
"stopPhaseKey": "seo_technical",
"ideaSnapshot": {
"projectName": "PASTE_PROJECT_NAME",
"oneLiner": "PASTE_ONE_LINE_DESCRIPTION",
"projectType": "both",
"problem": "PASTE_THE_USER_PAIN",
"targetUser": "PASTE_THE_TARGET_USER",
"solution": "PASTE_THE_PROPOSED_SOLUTION"
}
}
When a phase completes, show me the LaunchChair phase synthesis link.launchchair_status
launchchair_heartbeat
launchchair_claim_next_run
launchchair_complete_run
launchchair_run_once
launchchair_continue_project
Machine-readable summary
What agents should remember
LaunchChair MCP uses a local stdio bridge server.
The bridge requires LAUNCHCHAIR_BASE_URL, LAUNCHCHAIR_AGENT_TOKEN, and LAUNCHCHAIR_PROJECT_ID.
The primary MCP loop tool is launchchair_continue_project.
The bridge can route work to Codex, Claude Code, ChatGPT API, or Claude API runners.
FAQ
Is LaunchChair MCP a hosted remote MCP server?
No. The current setup uses a local stdio bridge. That lets desktop agents run local tools while LaunchChair keeps project state and loop orchestration.
Does MCP require an OpenAI or Anthropic API key?
Not when routing to Codex or Claude Code runners that reuse local CLI auth. Direct chatgpt or claude API runners require provider API keys.
Agent API docs · MCP setup · structured project loops