> ## Documentation Index
> Fetch the complete documentation index at: https://docs.lavendly.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Status & schema

> Two introspection endpoints. Use them at session start.

## `GET /v1/_status`

Capability flags, available voices and styles, current quotas. Hit
this at session start so the agent knows what works before it commits
credits.

```json theme={null}
{
  "ok": true,
  "capabilities": {
    "image":      true,
    "video":      true,
    "voice":      true,
    "music":      true,
    "transcribe": true,
    "storage":    true
  },
  "voice": {
    "supports_inline_tags": true,
    "available_styles":     ["natural", "narrator", "energetic", "calm"]
  },
  "models": {
    "video":  ["fast", "balanced", "cinematic"],
    "voice":  ["natural", "narrator", "energetic", "calm"],
    "music":  ["instrumental", "ambient", "score", "lofi"]
  }
}
```

Capability flags are coarse, `voice: true` means *something* can
synthesize a voiceover. The `models` lists tell you the catalog
without committing to a specific brand.

## `GET /v1/_schema`

The authoritative operation catalog. Every endpoint, its method, path,
body shape, return shape, and idempotency flag.

```json theme={null}
{
  "schema_version": 1,
  "base_url": "/v1",
  "auth": { "type": "bearer" },
  "error_envelope": {
    "shape": { "error": { "code": "string", "message": "string" } }
  },
  "operations": {
    "create_render": {
      "method":  "POST",
      "path":    "/workflows/{id}/renders",
      "body":    { "options": "object?" },
      "returns": "{ job_id, status, poll_url, ... }",
      "idempotent": true
    }
  }
}
```

Agents use this for **dynamic discovery**, the [MCP server's tool
catalog](/mcp/tools-reference) is generated from this same source.
When Lavendly adds a new operation, the schema endpoint reflects it
immediately and MCP clients that re-fetch the schema pick it up
without a code change.
