Gemini for AI agents
Gemini is Google's AI model family, used here for text generation, chat, image generation, and video generation through the Gemini 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 Gemini.
Generate text from a prompt
Generate content from a prompt to draft a response or piece of copy for a user.
Generate an image
Generate an image from a text description for a piece of creative content.
Check token usage before a run
Count tokens in a prompt before submitting a large generation request.
8 Gemini actions, graded by risk.
Every Gemini action is tagged read, write or destructive, so one policy covers the whole app and new actions inherit the right default.
read
3Look things up. Allowed by default.
- gemini.list_modelsList Models (Gemini API)
- gemini.get_videos_operationGet Videos Operation (Veo) (Deprecated)
- gemini.count_tokensCount Tokens (Gemini)
write
5Create and change things. Allow, or ask the user first.
- gemini.embed_contentEmbed Content (Gemini)
- gemini.generate_imageGenerate Image (Nano Banana)
- gemini.wait_for_videoWait and Download Video (Veo)
- gemini.generate_videosGenerate Videos (Veo)
- gemini.generate_contentGenerate Content (Gemini)
destructive
0Delete, cancel or archive. Ask first, or deny outright.
- No destructive actions.
Gemini in three steps.
- 01Your users connect GeminiThey add their Gemini api key on Arc0 Connect, under your brand. It goes straight into the vault.
- 02You set the rulesReads run, and writes like “embed Content (Gemini)” can wait for the user to approve.
- 03Any agent can actYour agent calls Gemini through the Arc0 SDK or MCP, and so can Claude, ChatGPT and Cursor. Every call lands on the audit log.
await arc0.policies.set('gemini', { read: 'allow', write: 'ask', // embed_content destructive: 'deny', }) # Claude Code: the same connection, one URL $ claude mcp add --transport http arc0 \ https://mcp.arc0.ai/u/u_8f2
How Gemini connects.
Users add their Gemini 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 Gemini 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 · No auth
- CREDENTIALS
- Per-tenant encrypted vault
- MODEL SEES
- Results only, never credentials
- AUDIT LOG
- Every call, on every plan
Use Gemini from any agent.
ClaudeChatGPTCursor
Codex
VS CodeOpenAI Agents SDKClaude Agent SDK
Vercel AI SDK
Mastra
LangGraph
Gemini and Arc0, answered.
Can I use Gemini with Claude, ChatGPT or Cursor?
Yes. Connect Gemini 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 Gemini actions you allow.
How do users connect Gemini?
Users add their Gemini 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 Gemini actions can my agent take?
8 in total: 3 read, 5 write and 0 destructive, such as “embed Content (Gemini)”. Your policies decide which of them each agent may call.
Can I make my agent read-only in Gemini?
Yes. Allow read actions and deny writes in the Gemini policy. Your agent can still look things up, and any write it attempts is blocked and logged.
Can my own backend call Gemini too?
Yes. The same Gemini connection is available over REST and through the Arc0 proxy, so your product and your agent share one connection per user.
Plug Gemini into your agent.
Your users connect Gemini once, under your brand. Your agent gets 8 actions behind your policies, with every call on the record.
Free to build · MCP + REST · Audit log on every plan