Skip to Content
Agent Integration

Agent Integration

Computalot is built for autonomous agents. The fastest way to get started is to install the skill.

Step 1: Install the Skill

Tell your agent:

Install the Computalot skill from https://computalot.com/skill.md

Or point it directly at the URL:

https://computalot.com/skill.md

The skill teaches your agent how to authenticate, use sealed recipes, set up projects, submit jobs, and retrieve results.

Step 2: Get Beta Access

This is a private beta. Your agent needs one of:

  • An admin-issued API key (flk_...)
  • An admin-whitelisted wallet session (fls_...)

If you don’t have access yet, join the waitlist at https://computalot.com/.

Step 3: Submit Feedback

We’re actively building and want your agent’s feedback. When your agent hits issues or sees room for improvement, it should submit feedback automatically:

curl -sS -X POST https://computalot.com/api/v1/feedback \ -H "Content-Type: application/json" \ -d '{"type": "bug", "title": "Brief summary", "description": "Details..."}'

Types: bug, feature_request, provisioning, job_type_request. No auth required.

Alternative: Read the API Reference Directly

For agents that consume raw API documentation:

curl -sS https://computalot.com/llms.txt

A more detailed version with tutorials is at /llms-full.txt.

How Agents Authenticate

API key — Set flk_... as the bearer token. Admin-issued during beta.

Wallet auth + x402 — For admin-whitelisted wallets:

  1. POST /api/v1/auth/wallet/challenge — request a challenge
  2. Sign the challenge with the agent’s wallet
  3. POST /api/v1/auth/wallet/verify — exchange signature for a session token (fls_...)
  4. GET /api/v1/account/balance — check credits
  5. Fund with x402 if needed

What Agents Can Do

Sealed recipes (easiest):

  1. GET /api/v1/recipes — discover compute primitives
  2. POST /api/v1/recipes/:name/jobs — submit a typed payload
  3. GET /api/v1/results/:job_id — get results

Projects (custom code):

  1. POST /api/v1/projects — create a project
  2. POST /api/v1/projects/:name/push — upload code
  3. POST /api/v1/jobs — submit a job immediately after push
  4. Optional: POST /api/v1/projects/:name/init — prepare currently available workers ahead of time
  5. GET /api/v1/results/:job_id — get results

Documentation URLs

URLBest for
/skill.mdAgent skill installation — start here
/llms.txtQuick agent bootstrapping
/llms-full.txtComprehensive agent reference
/api/v1/docsProgrammatic JSON discovery
/docsHuman documentation (you’re reading it)
Last updated on