Skip to main content
Use AutoAdy from a script, an agent, or a terminal. All three interfaces act with your AutoAdy identity and check account access, tool availability, and plan requirements.

Quickstart

  1. Create a named API key. For a Meta read, first connect Meta with the intended identity.
  2. Put the key in your private runtime environment as AUTOADY_API_KEY.
  3. List the accounts that identity can reach:
  4. Inspect data.accounts. Use a returned account ID in your first performance read:
The ID above is a placeholder. A successful empty account list proves authentication, but does not prove a reachable Meta connection. Check the intended identity, connection, and workspace grants before continuing.

Authentication and errors

REST uses Authorization: Bearer adk_.... OAuth-capable MCP clients can use account sign-in instead. Authentication covers key creation, expiry, permissions, and revocation. Requests and responses explains both error envelopes and recovery.

Endpoint reference

The REST reference groups the 52 currently published POST endpoints by task. MCP has a wider catalog; use its tools/list response for the actual tools offered by that server. Tool names, argument shapes, and availability can differ between the two interfaces.

MCP server

The remote MCP endpoint remains https://www.autoady.io/api/mcp. Follow MCP and CLI setup, or the client setup guide for a first verified read.

Command line

Use npx -y @autoady/cli or install the CLI globally. Command line setup covers environment credentials, commands, raw JSON calls, and exit codes.

OpenAPI specification

The public OpenAPI 3.1 specification supplies request schemas, examples, and response definitions. Fetch it without authentication to configure your HTTP client. Keep credentials out of generated files and shared examples. Start with reads. Before any account change, review its exact target and proposed effect with the user, then follow the tool’s returned confirmation and retry contract.