Browserbase for AI agents
Browserbase gives AI agents isolated cloud browsers that developers use to navigate sites, extract data, and run natural-language web actions at scale. 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 Browserbase.
Start an isolated browser session
Start a cloud browser session so an agent can navigate a site without touching the user's own browser.
Extract data from a page
Navigate to a page and extract structured data after the user confirms which site to visit.
Require approval before an action step
Treat a natural-language act step that submits a form or clicks a button as needing confirmation before it runs.
6 Browserbase actions, graded by risk.
Every Browserbase action is tagged read, write or destructive, so one policy covers the whole app and new actions inherit the right default.
read
1Look things up. Allowed by default.
- browserbase.extractExtract
write
5Create and change things. Allow, or ask the user first.
- browserbase.actAct
- browserbase.endEnd
- browserbase.startStart
- browserbase.observeObserve
- browserbase.navigateNavigate
destructive
0Delete, cancel or archive. Ask first, or deny outright.
- No destructive actions.
Browserbase in three steps.
- 01Your users connect BrowserbaseThey add their Browserbase api key on Arc0 Connect, under your brand. It goes straight into the vault.
- 02You set the rulesReads run, and writes like “act” can wait for the user to approve.
- 03Any agent can actYour agent calls Browserbase through the Arc0 SDK or MCP, and so can Claude, ChatGPT and Cursor. Every call lands on the audit log.
await arc0.policies.set('browserbase', { read: 'allow', write: 'ask', // act destructive: 'deny', }) # Claude Code: the same connection, one URL $ claude mcp add --transport http arc0 \ https://mcp.arc0.ai/u/u_8f2
How Browserbase connects.
Users add their Browserbase 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 Browserbase 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 Browserbase from any agent.
ClaudeChatGPTCursor
Codex
VS CodeOpenAI Agents SDKClaude Agent SDK
Vercel AI SDK
Mastra
LangGraph
Browserbase and Arc0, answered.
Can I use Browserbase with Claude, ChatGPT or Cursor?
Yes. Connect Browserbase 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 Browserbase actions you allow.
How do users connect Browserbase?
Users add their Browserbase 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 Browserbase actions can my agent take?
6 in total: 1 read, 5 write and 0 destructive, such as “act”. Your policies decide which of them each agent may call.
Can I make my agent read-only in Browserbase?
Yes. Allow read actions and deny writes in the Browserbase policy. Your agent can still look things up, and any write it attempts is blocked and logged.
Can my own backend call Browserbase too?
Yes. The same Browserbase connection is available over REST and through the Arc0 proxy, so your product and your agent share one connection per user.
Plug Browserbase into your agent.
Your users connect Browserbase once, under your brand. Your agent gets 6 actions behind your policies, with every call on the record.
Free to build · MCP + REST · Audit log on every plan