# ORDan agent documentation

ORDan gives autonomous agents a communications identity: dedicated phone
numbers, voice, messaging, human-delegated permissions, and an auditable
record — through one API. Your runtime supplies the intelligence.

**Status:** ORDan is in staged deployment with paid billing under live
verification. Use your operator's API URL (local default
`http://localhost:4000`). The SDK, CLI and MCP packages are prepared for
npm and PyPI but not yet published. Telephone and messaging delivery depend
on carrier networks; messaging campaign registration is pending carrier
verification.

## Start here

- [Quickstart](quickstart.md): pair an agent or use an owner-authorized project key.
- [Authentication and authorization](authentication.md): roles, scopes, grants, credentials.
- [API conventions and reference](api.md): HTTP, schemas, identifiers and units.
- [Lines and numbers](numbers.md): asynchronous acquisition, holds and release.
- [Voice, recordings and transfers](voice.md): lifecycle, budgets and external runtimes.
- [Messaging](messaging.md): segments, carrier fees and 10DLC registration.
- [Errors and retries](errors.md): approvals, idempotency and unknown outcomes.
- [Webhooks and events](webhooks.md): signed deliveries, replay and event polling.
- [SDK, CLI and MCP](clients.md): client setup, package names and human pairing.

The running API exposes its contract at `GET /v1/openapi.json`.

Legal: [Privacy Policy](https://ordan.io/privacy) ·
[Terms of Service](https://ordan.io/terms)