Are you an LLM? Read llms.txt for a summary of the docs, or llms-full.txt for the full context.
Skip to content

Build With A Coding Agent

Paste this into your coding agent:

Fetch this setup prompt and follow it to build and deploy my Amodal agent:
https://docs.amodalai.com/agent-setup/prompt.md

The prompt tells the agent to download these docs as text (llms-full.txt) and the Platform API reference (openapi.json), collect missing configuration, connect source, deploy, and test the result.

Platform Token

The coding agent needs a pk_ platform automation token to create, configure, or deploy agents:

  1. Open Amodal and select your organization.
  2. Open Organization > Automation Tokens.
  3. Create an org-wide admin token for initial setup.

Creating an agent requires an admin token. Once the agent exists, an agent-bound operator token can deploy, activate, promote, roll back, and mint runtime tokens. An observer token can only read. Automation tokens cannot create or rotate other automation tokens.

Also give the coding agent the organization ID, the source repository, and any runtime secrets the source references.

A platform token does not authorize chat. Chat smoke tests use a short-lived runtime token minted with POST /api/agents/{agent_id}/tokens; see Runtime Server.