Deepgram
Also an agent toolDeepgram voice AI REST API: transcribe pre-recorded audio from a URL (speech-to-text), synthesize speech from text (text-to-speech), and manage projects, API keys, members, invitations, usage, balances and models.
26 actions · 0 triggers
The Deepgram integration lets you connect Deepgram to your automations with 26 actions — and every action is also a tool an AI agent can call. Examples: List Projects, Transcribe Audio URL, Synthesize Speech (Text-to-Speech), Get Project, Update Project. No blank-canvas wiring — connect it and use it in your workflows.
Every action below is also a tool an AI agent can call — the same pack powers your workflows and your agents.
Actions 26
What Zervanor can do in Deepgram.
Projects22
List Projects
List the projects associated with the API key (GET /projects). Identity + connectivity probe.
Get Project
Get details for a specific project (GET /projects/{project_id}).
Update Project
Update a project's name and settings (PATCH /projects/{project_id}).
Delete Project
Delete a project (DELETE /projects/{project_id}).
List Project Models
List the models a project has access to (GET /projects/{project_id}/models).
Get Project Model
Get metadata for a specific model in a project (GET /projects/{project_id}/models/{model_id}).
List Project Keys
List the API keys for a project (GET /projects/{project_id}/keys).
Get Project Key
Get details for a specific API key (GET /projects/{project_id}/keys/{key_id}).
Create Project Key
Create an API key for a project (POST /projects/{project_id}/keys).
Delete Project Key
Delete an API key from a project (DELETE /projects/{project_id}/keys/{key_id}).
List Project Members
List the members of a project (GET /projects/{project_id}/members).
Remove Project Member
Remove a member from a project (DELETE /projects/{project_id}/members/{member_id}).
List Member Scopes
List the scopes assigned to a project member (GET /projects/{project_id}/members/{member_id}/scopes).
Update Member Scopes
Update the scope assigned to a project member (PUT /projects/{project_id}/members/{member_id}/scopes).
List Invitations
List the pending invitations for a project (GET /projects/{project_id}/invites).
Create Invitation
Invite a person to a project by email (POST /projects/{project_id}/invites).
Delete Invitation
Remove a pending invitation by email (DELETE /projects/{project_id}/invites/{email}).
Get Usage Summary
Get aggregated usage for a project over a date range (GET /projects/{project_id}/usage).
Get Usage Breakdown
Get a grouped usage breakdown for a project (GET /projects/{project_id}/usage/breakdown).
Get Usage Fields
List the tags, models, methods and features seen in a project's requests (GET /projects/{project_id}/usage/fields).
List Balances
List the outstanding balances for a project (GET /projects/{project_id}/balances).
Get Balance
Get details for a specific balance (GET /projects/{project_id}/balances/{balance_id}).
Models2
List Models
List all latest STT and TTS models available (GET /models).
Get Model
Get metadata for a specific model (GET /models/{model_id}).
Listen1
Transcribe Audio URL
Transcribe pre-recorded audio hosted at a public URL (POST /listen). The audio location is passed as a JSON body { url }; transcription options are query parameters. Supports an async callback URL for long jobs.
Speak1
Synthesize Speech (Text-to-Speech)
Convert text into natural-sounding speech (POST /speak). The text is passed as a JSON body { text }; voice and audio options are query parameters. The response is binary audio.
Nothing matches your search.
Automations using Deepgram
Build your own automation with Deepgram in the visual builder.
Frequently asked questions
What can I automate with Deepgram?
26 actions are available — for example: List Projects, Transcribe Audio URL, Synthesize Speech (Text-to-Speech), Get Project, Update Project, and more.
Can Deepgram be used as an AI agent tool?
Yes. Every Deepgram action automatically becomes a tool an AI agent can call — the same pack powers both your workflows and your agents.
How is the Deepgram integration modelled?
Deepgram is modelled as a versioned connector pack, and the steps that need reasoning fall to first-party agents hosted in the same runtime as the workflows.