Hatchet MCP runs on your own machine — the client starts it, so there is no endpoint to ping. 39 installs a week from npm. Last commit 31 May 2026.
Observe and operate Hatchet workflows from an AI agent — runs, logs, trigger, cancel, replay.
Nothing serious here today
Today is the operative word: we check Hatchet MCP every 15 minutes and re-read its code on every release. Watch it and you find out the day that stops being true.
Connect this server
This server runs on your own machine — install it with the package manager and the client starts it for you. Package name taken from the official registry entry.
claude mcp add hatchet-mcp -e 'HATCHET_CLIENT_TOKEN=<HATCHET_CLIENT_TOKEN>' -- npx -y hatchet-mcp{
"mcpServers": {
"hatchet-mcp": {
"command": "npx",
"args": [
"-y",
"hatchet-mcp"
],
"env": {
"HATCHET_CLIENT_TOKEN": "<HATCHET_CLIENT_TOKEN>"
}
}
}
}[mcp_servers.hatchet-mcp]
command = "npx"
args = ["-y", "hatchet-mcp"]
env = { HATCHET_CLIENT_TOKEN = "<HATCHET_CLIENT_TOKEN>" }{
"mcpServers": {
"hatchet-mcp": {
"command": "npx",
"args": [
"-y",
"hatchet-mcp"
],
"env": {
"HATCHET_CLIENT_TOKEN": "<HATCHET_CLIENT_TOKEN>"
}
}
}
}{
"servers": {
"hatchet-mcp": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"hatchet-mcp"
],
"env": {
"HATCHET_CLIENT_TOKEN": "<HATCHET_CLIENT_TOKEN>"
}
}
}
}Put your own value in place of
<HATCHET_CLIENT_TOKEN> (Hatchet API token (a JWT). Encodes the server URL and tenant, so it is normally the only setting required).
The author’s registry entry marks it as required.
Optional settings from the registry entry, left out of the block above:
HATCHET_API_BASE (Override the API base URL. Self-hosters can point this at any Hatchet instance), HATCHET_TENANT_ID (Override the tenant id decoded from the token).
Alternatives to Hatchet MCP
same job, measured the same wayCost per run, tool, model and day for AI agents, from one call or an OpenTelemetry export.
Read a tool-call log from an AI agent run and see which calls failed, looped or were slow.
Airflow MCP — list DAGs/runs/task instances, tail logs, trigger and clear (write-gated)
Read-only MCP for OpenVibeCoding runs, workflows, proof, queue, replay, and incident reads.
CircleCI's MCP server: inspect projects, runs, workflows, jobs, logs, tests, and artifacts.
Errors, traces, replays and logs from Talaria for coding agents.
Measures the cost, tokens and duration of a completed agent workflow from its session log.
Step-by-step observability for MCP agent workflows
Hatchet MCP — questions
Answers built from our own checks of this server.