ADLAgent Definition Language

Why ADL? ​

The AI agent ecosystem is fragmenting fast. Every provider has a different surface; every framework has its own scaffolding. ADL gives teams one declarative manifest from which enterprise-ready agent code, configuration, documentation, and deployment manifests can be generated - vendor-agnostic and portable across platforms.

apiVersion: adl.inference-gateway.com/v1
kind: Agent
metadata:
  name: customer-support
  description: AI agent for handling customer inquiries
  version: "1.0.0"
spec:
  capabilities:
    streaming: true
    pushNotifications: true
    stateTransitionHistory: true
  agent:
    provider: deepseek
    model: deepseek-v4-flash
    systemPrompt: You are a professional customer support agent.
  server:
    port: 8080
  language:
    go:
      module: github.com/company/customer-support-agent
      version: "1.26"

Feed that file into adl-cli and get a complete agent project - server, handlers, tests, sandbox, CI, and deployment manifests - in the language of your choice. The Generate & Run guide walks the whole path, from install to a server answering on localhost.

Where to go next ​

• New to ADL? Start with What is ADL? and then Getting Started.
• Want a running agent? Follow Generate & Run from manifest to localhost.
• Looking up a specific field? Jump into the Schema Reference.
• Want to see real manifests? Browse the Examples.