Deepgram for AI agents
Deepgram provides AI speech recognition and text-to-speech, letting a product transcribe recorded audio or generate spoken responses through one API. Connect it once through Arc0, and your agent, or Claude, ChatGPT and Cursor, can use it through one MCP endpoint, limited to what each user approved.
What agents do in Deepgram.
Transcribe a recorded file
Transcribe a pre-recorded audio file into text for search, review, or analysis.
Generate speech from text
Generate a spoken audio clip from text using a chosen voice model.
Check project usage
Check a project's usage summary before it runs into its balance limit.
17 Deepgram actions, graded by risk.
Every Deepgram action is tagged read, write or destructive, so one policy covers the whole app and new actions inherit the right default.
read
15Look things up. Allowed by default.
- deepgram.get_modelGet Model by ID
- deepgram.get_modelsGet Public Models
- deepgram.get_projectsList Deepgram Projects
- deepgram.list_think_modelsList Think Models
- deepgram.get_project_balanceGet Project Balance
- deepgram.get_project_requestGet Project Request
- deepgram.list_project_scopesList Project Scopes
- deepgram.list_project_balancesList Project Balances
- deepgram.list_project_requestsList Project Requests
- deepgram.list_project_purchasesList Project Purchases
- deepgram.get_project_usage_summaryGet Project Usage Summary
- deepgram.list_project_usage_fieldsList Project Usage Fields
- deepgram.get_project_usage_breakdownGet Project Usage Breakdown
- deepgram.list_project_billing_fieldsList Project Billing Fields
- deepgram.get_project_billing_breakdownGet Project Billing Breakdown
write
2Create and change things. Allow, or ask the user first.
- deepgram.text_to_speech_restText-to-Speech (REST)
- deepgram.speech_to_text_pre_recordedTranscribe Pre-recorded Audio
destructive
0Delete, cancel or archive. Ask first, or deny outright.
- No destructive actions.
Deepgram in three steps.
- 01Your users connect DeepgramThey add their Deepgram api key on Arc0 Connect, under your brand. It goes straight into the vault.
- 02You set the rulesReads run, and writes like “text-to-Speech (REST)” can wait for the user to approve.
- 03Any agent can actYour agent calls Deepgram through the Arc0 SDK or MCP, and so can Claude, ChatGPT and Cursor. Every call lands on the audit log.
await arc0.policies.set('deepgram', { read: 'allow', write: 'ask', // text_to_speech_rest destructive: 'deny', }) # Claude Code: the same connection, one URL $ claude mcp add --transport http arc0 \ https://mcp.arc0.ai/u/u_8f2
How Deepgram connects.
Users add their Deepgram api key on Arc0 Connect. It is encrypted in the vault, never shown to the model, and each user can rotate or revoke it at any time.
The same Deepgram connection serves your agent over MCP and your own backend over REST and the proxy, so a user connects once. How Arc0 handles credentials →
- AUTH
- API key
- CREDENTIALS
- Per-tenant encrypted vault
- MODEL SEES
- Results only, never credentials
- AUDIT LOG
- Every call, on every plan
Use Deepgram from any agent.
Deepgram and Arc0, answered.
Can I use Deepgram with Claude, ChatGPT or Cursor?
Yes. Connect Deepgram to Arc0 once, then add your Arc0 MCP URL to Claude, ChatGPT, Cursor, Claude Code or any other remote-MCP client. Each assistant only gets the Deepgram actions you allow.
How do users connect Deepgram?
Users add their Deepgram api key on Arc0 Connect. It is encrypted in the vault, never shown to the model, and each user can rotate or revoke it at any time.
Which Deepgram actions can my agent take?
17 in total: 15 read, 2 write and 0 destructive, such as “text-to-Speech (REST)”. Your policies decide which of them each agent may call.
Can I make my agent read-only in Deepgram?
Yes. Allow read actions and deny writes in the Deepgram policy. Your agent can still look things up, and any write it attempts is blocked and logged.
Can my own backend call Deepgram too?
Yes. The same Deepgram connection is available over REST and through the Arc0 proxy, so your product and your agent share one connection per user.
Plug Deepgram into your agent.
Your users connect Deepgram once, under your brand. Your agent gets 17 actions behind your policies, with every call on the record.
Free to build · MCP + REST · Audit log on every plan