Hatchet MCP

@ElliotPadfield

About Hatchet MCP

MCP server for Hatchet — let AI agents observe and operate your workflows: runs, logs, workers, metrics, plus trigger/cancel/replay. Single-token setup, npx-installable.

Config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "hatchet": {
      "command": "npx",
      "args": [
        "-y",
        "hatchet-mcp"
      ],
      "env": {
        "HATCHET_CLIENT_TOKEN": "<your-hatchet-api-token>"
      }
    }
  }
}

Tools

No tools detected

Fetch the live tool list by running this server in a temporary sandbox using the button above.

Overview

What is Hatchet MCP?

Hatchet MCP is a server that allows AI agents, such as Claude Code or Claude Desktop, to observe and operate Hatchet workflows. It wraps the Hatchet API to provide status, runs, logs, workers, and metrics, plus the ability to trigger, cancel, and replay runs.

How to use Hatchet MCP?

Add the server to your MCP configuration using npx -y hatchet-mcp with the required environment variable HATCHET_CLIENT_TOKEN (obtained from the Hatchet dashboard under API tokens). Optionally set HATCHET_API_BASE for self-hosted instances and HATCHET_TENANT_ID to override the tenant decoded from the token. Once configured, agents can invoke the provided tools to interact with Hatchet workflows.

Key features of Hatchet MCP

Read-only observability tools for workflows, runs, logs, workers, and queue metrics
Action tools to trigger, cancel, and replay workflow runs
Single environment variable (HATCHET_CLIENT_TOKEN) encodes server URL and tenant
Works with Hatchet Cloud and self-hosted instances via HATCHET_API_BASE
Built with TypeScript and ESM, tested with vitest

Use cases of Hatchet MCP

Monitor workflow run status, tasks, and errors via an AI assistant
Trigger new workflow runs by name with a JSON input payload
Cancel or replay one or more runs or tasks by external ID
Inspect worker health and queue metrics to assess system load
Debug tasks by fetching log lines during incident response

FAQ from Hatchet MCP

How do I get the HATCHET_CLIENT_TOKEN?

Obtain the token from the Hatchet dashboard under API tokens. It is a JWT that encodes the server URL and tenant, making it the only required setting.

Can I use Hatchet MCP with a self-hosted Hatchet instance?

Yes. Set the HATCHET_API_BASE environment variable to point at your own Hatchet instance and the server will work anywhere.

Which tools modify live state?

The tools trigger_workflow, cancel_runs, and replay_runs mutate live state. Their descriptions are prefixed with “MUTATES LIVE STATE” so agents and users know they affect real runs.

Is the token safe to share or commit?

No. The token grants full tenant access; treat it as a secret and never commit it to source control.

What is the current status of the package?

Version 0.1.0. All tools have been verified against Hatchet Cloud and work with self-hosted instances. The trigger_workflow tool uses the stable /workflow-runs/trigger endpoint.

Frequently asked questions

How do I get the HATCHET_CLIENT_TOKEN?

Obtain the token from the Hatchet dashboard under API tokens. It is a JWT that encodes the server URL and tenant, making it the only required setting.

Can I use Hatchet MCP with a self-hosted Hatchet instance?

Yes. Set the `HATCHET_API_BASE` environment variable to point at your own Hatchet instance and the server will work anywhere.

Which tools modify live state?

The tools `trigger_workflow`, `cancel_runs`, and `replay_runs` mutate live state. Their descriptions are prefixed with “MUTATES LIVE STATE” so agents and users know they affect real runs.

Is the token safe to share or commit?

No. The token grants full tenant access; treat it as a secret and never commit it to source control.

What is the current status of the package?

Version 0.1.0. All tools have been verified against Hatchet Cloud and work with self-hosted instances. The `trigger_workflow` tool uses the stable `/workflow-runs/trigger` endpoint.

Comments

Basic information

More Developer Tools MCP servers

TaskerArmy Agent

TaskerArmy

Ask Claude or ChatGPT what Shopify theme optimization tasks are pending on your store, a remote MCP server for TaskerArmy Agent accounts.

Reelier

Maxime Houle

Agents make claims. Reelier writes receipts — record an agent's tool-call workflow once, replay it deterministically at 0 tokens, and diff runs to catch drift.

AptiBuild AI — Career Intelligence

ParklandBuilds

AptiBuild AI — Career Intelligence MCP Server Structured career decision scaffolding for AI agents. Provides personality-conditioned business idea generation, official BLS OEWS wage data, FRED labor market signals, and

Raccha AI

Shree Mandadi

MCP-first toolbox for agents: namespaced key-value storage, a FIFO queue, auth, and a handful of stateless utility tools (JWT decode, hashing, cert inspection, CIDR math). Free while in early access. This repository is a

TranscriptFetch MCP Server

TranscriptFetch

Model Context Protocol (MCP) server for TranscriptFetch: fetch YouTube transcripts, search, channels, and playlists from any MCP client.