Deepgram
AI-powered speech-to-text, text-to-speech, and audio intelligence API for transcribing audio, synthesizing speech, and managing voice AI projects.
Three ways to use Deepgram
Chat with Geni
Ask Geni to pull data, take actions, or answer questions using this integration in a conversation.
Build a workflow
Create automated workflows that trigger on events, run on a schedule, or chain multiple tools together.
Power an app
Use this integration as a data source or action layer behind a dashboard, form, or internal tool.
Supported tools
Actions your AI agents can perform with Deepgram.
Get Model by ID
Retrieve metadata for a single Deepgram model (STT or TTS) by UUID.
Get Project Usage Summary
Get aggregated usage stats for a Deepgram project: audio duration, billable duration, request count, and quality metrics, with breakdowns by model/tag/accessor.
List Projects
List all Deepgram projects accessible to the authenticated API key.
List Project Scopes
List the permission scopes available on a Deepgram project (member, admin, owner role assignments).
List Public Models
List all public Deepgram speech-to-text and text-to-speech models with metadata, supported languages, and capabilities.
List Think Models
List the LLM "think" models available for use inside Deepgram Voice Agent configurations.
Get started automating Deepgram
Ready-to-use workflows
Prompts using Deepgram
Call overdue Xero customers with an AI collections agent
Every weekday at 10am, an AI voice agent phones your most overdue Xero accounts, logs what each customer promised, and reports back to finance in Slack.
Weekly sales call coaching board with rubric scorecards
Pull this week's call recordings into one board, score every rep against the same rubric, and see at a glance who still needs a review.
Interview debrief desk that gets feedback written the same day
See every interview still missing a writeup, listen back to what the candidate actually said, and file the feedback before you forget it.
Turn every WordPress post into a narrated audio version
A studio your content team opens after publishing to clean up the script, pick a voice, and ship a listenable version of every post.
Turn LiveKit call recordings into HubSpot notes and Slack recaps
Every 30 minutes, an agent transcribes finished LiveKit calls, files a clean summary and action items on the right HubSpot contact, and posts a recap to your sales channel.
Triage support call recordings into Zendesk every morning
Every weekday at 8am, transcribe yesterday's support calls, add a clean summary to the matching Zendesk ticket, and flag negative ones for your support lead.
Turn user interview recordings into an Airtable insight repository
Every Friday at 4pm, transcribe this week's user interviews from Google Drive and file the most quotable moments as themed rows in your Airtable research repository.
Turn morning voice memos into Linear issues every weekday
Every weekday at 9am, transcribe yesterday's walk and talk voice memos from Dropbox and file each action item as a clean Linear issue for your team.