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

# Deepgram

> Connect Deepgram to your Atako agents — 6 read and 2 write actions.

Let your agents transcribe audio from URLs, analyse text with AI (summarise, topics, intents, sentiment), and inspect Deepgram projects, usage and billing.

## Connection

* **Authentication**: API key.

<Note>
  console.deepgram.com → sign in → API Keys → Create a New API Key. Choose a project, set the scopes to "Member" (or custom: usage:read + usage:read/write), and copy the key.

  See [Deepgram's documentation](https://console.deepgram.com/signup).
</Note>

## Read actions (6)

| Action | Description |
| - | - |
| `get_project` | Fetch details of a specific project by its ID. |
| `get_project_balances` | List outstanding balances for a project. |
| `get_project_usage` | Get usage summary for a project. Pass start and end as YYYY-MM-DD dates to filter the range. |
| `list_models` | List all available public models (STT, TTS, etc.). |
| `list_project_keys` | List API keys for a project. Key values are never returned — only IDs, comments, scopes and creation dates. |
| `list_projects` | List all projects accessible with the API key. |

## Write actions (2)

| Action | Description |
| - | - |
| `text_intelligence` | Analyse plain text with Deepgram's text intelligence API. Body: \{ "text": "…" }. Query params: summarize (boolean — generate a summary), topics (boolean — detect topics), intents (boolean — recognise intent), sentiment (boolean — analyse sentiment), language (BCP-47 tag). All query params are optional. |
| `transcribe_url` | Transcribe audio or video from a URL. Body: \{ "url": "https\://…" }. Query params: model (e.g. "nova-3"), language (BCP-47 tag, e.g. "en"), smart\_format (boolean — apply extra formatting), diarize\_model (string — "latest", "v1" or "v2" — recognise speaker changes), punctuate (boolean — add punctuation), summarize (boolean or string — summarise the transcript), detect\_language (boolean — identify dominant language), paragraphs (boolean — split into paragraphs), utterances (boolean — segment into semantic units). All query params are optional. |

## Permissions

Every action above must be explicitly granted to an agent before it can be used. See [Permissions](/integrations/permissions) for the grant model and [Security](/integrations/security) for how credentials are protected.

***

*Last reviewed against the provider API: September 2026.*


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.