Gradium for AI agents

Gradium provides low-latency speech generation, transcription, voice management, and pronunciation tools that developers use to build real-time voice applications like assistants and interactive phone systems. 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.

AI and modelsAPI keyMCP + RESTgradium.ai
AUDIT LOG · GRADIUMPOLICY: acme-support
09:41:07 · claude · u_8f2read
gradium.list_voices
List Voices✓ allowed · 212ms
09:41:08 · claude · u_8f2read
gradium.get_voice
Get Voice✓ allowed · 164ms
09:41:09 · claude · u_8f2write
gradium.create_pronunciation_dictionary
Create Pronunciation Dictionary✓ approved · approved by user
09:41:12 · claude · u_8f2destructive
gradium.delete_pronunciation_dictionary
Delete Pronunciation Dictionary✕ blocked · policy: deny
EVERY GRADIUM CALL, ON THE RECORD
01 · USE CASES

What agents do in Gradium.

01

List available voices

List the voices available on the account before picking one for a new voice assistant.

02

Check usage and credits

Pull the account's usage summary to see remaining credits before a large synthesis job.

03

Update a pronunciation dictionary

Update a custom pronunciation entry so a brand name is read correctly in generated speech.

02 · ACTIONS

9 Gradium actions, graded by risk.

Every Gradium action is tagged read, write or destructive, so one policy covers the whole app and new actions inherit the right default.

read

6

Look things up. Allowed by default.

  • gradium.get_voice
    Get Voice
  • gradium.get_credits
    Get Credits
  • gradium.list_voices
    List Voices
  • gradium.get_usage_summary
    Get Usage Summary
  • gradium.get_pronunciation_dictionary
    Get Pronunciation Dictionary
  • gradium.list_pronunciation_dictionaries
    List Pronunciation Dictionaries

write

2

Create and change things. Allow, or ask the user first.

  • gradium.create_pronunciation_dictionary
    Create Pronunciation Dictionary
  • gradium.update_pronunciation_dictionary
    Update Pronunciation Dictionary

destructive

1

Delete, cancel or archive. Ask first, or deny outright.

  • gradium.delete_pronunciation_dictionary
    Delete Pronunciation Dictionary
03 · HOW IT WORKS

Gradium in three steps.

  1. 01Your users connect GradiumThey add their Gradium api key on Arc0 Connect, under your brand. It goes straight into the vault.
  2. 02You set the rulesReads run, writes like “create Pronunciation Dictionary” can wait for the user, and “delete Pronunciation Dictionary” can be denied outright.
  3. 03Any agent can actYour agent calls Gradium through the Arc0 SDK or MCP, and so can Claude, ChatGPT and Cursor. Every call lands on the audit log.
POLICY.TS
await arc0.policies.set('gradium', {
  read: 'allow',
  write: 'ask',        // create_pronunciation_dictionary
  destructive: 'deny',  // delete_pronunciation_dictionary
})

# Claude Code: the same connection, one URL
$ claude mcp add --transport http arc0 \
    https://mcp.arc0.ai/u/u_8f2
04 · AUTH AND DATA

How Gradium connects.

Users add their Gradium 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 Gradium 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
05 · WORKS WITH

Use Gradium from any agent.

Claude
ChatGPT
Cursor
Codex
VS Code
OpenAI Agents SDK
Claude Agent SDK
Vercel AI SDK
Mastra
LangGraph
07 · FAQ

Gradium and Arc0, answered.

Q01

Can I use Gradium with Claude, ChatGPT or Cursor?

Yes. Connect Gradium 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 Gradium actions you allow.

Q02

How do users connect Gradium?

Users add their Gradium 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.

Q03

Which Gradium actions can my agent take?

9 in total: 6 read, 2 write and 1 destructive, such as “create Pronunciation Dictionary”. Your policies decide which of them each agent may call.

Q04

Can I stop my agent from deleting things in Gradium?

Yes. Actions like “delete Pronunciation Dictionary” are graded destructive. Set destructive actions to deny, or to ask so the user approves each one, and blocked calls still show up on the audit log.

Q05

Can my own backend call Gradium too?

Yes. The same Gradium connection is available over REST and through the Arc0 proxy, so your product and your agent share one connection per user.

Get started

Plug Gradium into your agent.

Your users connect Gradium once, under your brand. Your agent gets 9 actions behind your policies, with every call on the record.

Free to build · MCP + REST · Audit log on every plan