Spoolis

Coding-agent integration

Use Spoolis from any coding agent.

An agent can report that the work is done. Before another agent, deploy pipeline, billing system, or marketplace acts on that report, it should be able to check whether the work met the criteria everyone agreed to. Spoolis turns that result into a signed Outcome Receipt that the next system can verify.

MCP-capable agents

Connect to the hosted MCP endpoint.

Use this path for a Devin-style autonomous coding agent or any client that accepts a Streamable HTTP MCP server. No key is needed for sandbox tools. Production calls use a Spoolis API key.

MCP configuration
{
  "mcpServers": {
    "spoolis": {
      "url": "https://spoolis.com/api/mcp"
    }
  }
}

Claude Code and Codex

Run the npm MCP server over stdio.

This config works without a key against the sandbox. Add the documented environment block when you are ready to use an authenticated production key.

.mcp.json
{
  "mcpServers": {
    "spoolis": {
      "command": "npx",
      "args": ["-y", "@spoolis/mcp"]
    }
  }
}

Try it with no key

Ask the agent to run a sandbox scenario.

After connecting, call run_sandbox_scenario. The sandbox lets an agent inspect the complete flow without an account or production credential. Sandbox artifacts are demo-only and do not authorize production settlement or move funds.

Keep going

Read the contracts behind the integration.

Coding-agent quickstartRun a complete sandbox verification.MCP server guideSee tools, transports, and authenticated setup.Outcome Receipt specificationUnderstand what the signed result proves.Hugging Face agentsVerify delegated work before the manager acts.The agent said doneSee why the next system still needs a checkable result.
Use Spoolis from any coding agent