DeepWiki for AI agents
DeepWiki gives free AI-powered access to public GitHub repository documentation, letting a developer ask questions about a codebase instead of reading it all. 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 DeepWiki.
Ask a question about a repository
Ask a question about a public repository's structure instead of reading through the code.
Read a wiki's structure
Read a repository's generated wiki structure to see what documentation already exists.
Read specific wiki contents
Read a specific section of a repository's wiki for details on a subsystem.
3 DeepWiki actions, graded by risk.
Every DeepWiki action is tagged read, write or destructive, so one policy covers the whole app and new actions inherit the right default.
read
2Look things up. Allowed by default.
- deepwiki.read_wiki_contentsRead wiki contents
- deepwiki.read_wiki_structureRead wiki structure
write
1Create and change things. Allow, or ask the user first.
- deepwiki.ask_wiki_questionAsk wiki question
destructive
0Delete, cancel or archive. Ask first, or deny outright.
- No destructive actions.
DeepWiki in three steps.
- 01Your users connect DeepWikiThey add their DeepWiki no auth on Arc0 Connect, under your brand. It goes straight into the vault.
- 02You set the rulesReads run, and writes like “ask wiki question” can wait for the user to approve.
- 03Any agent can actYour agent calls DeepWiki through the Arc0 SDK or MCP, and so can Claude, ChatGPT and Cursor. Every call lands on the audit log.
await arc0.policies.set('deepwiki', { read: 'allow', write: 'ask', // ask_wiki_question destructive: 'deny', }) # Claude Code: the same connection, one URL $ claude mcp add --transport http arc0 \ https://mcp.arc0.ai/u/u_8f2
How DeepWiki connects.
DeepWiki needs no credentials. Arc0 still applies your policies and records every call, so public tools get the same controls as private ones.
The same DeepWiki 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
- No auth
- CREDENTIALS
- Per-tenant encrypted vault
- MODEL SEES
- Results only, never credentials
- AUDIT LOG
- Every call, on every plan
Use DeepWiki from any agent.
ClaudeChatGPTCursor
Codex
VS CodeOpenAI Agents SDKClaude Agent SDK
Vercel AI SDK
Mastra
LangGraph
DeepWiki and Arc0, answered.
Can I use DeepWiki with Claude, ChatGPT or Cursor?
Yes. Connect DeepWiki 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 DeepWiki actions you allow.
How do users connect DeepWiki?
DeepWiki needs no credentials. Arc0 still applies your policies and records every call, so public tools get the same controls as private ones.
Which DeepWiki actions can my agent take?
3 in total: 2 read, 1 write and 0 destructive, such as “ask wiki question”. Your policies decide which of them each agent may call.
Can I make my agent read-only in DeepWiki?
Yes. Allow read actions and deny writes in the DeepWiki policy. Your agent can still look things up, and any write it attempts is blocked and logged.
Can my own backend call DeepWiki too?
Yes. The same DeepWiki connection is available over REST and through the Arc0 proxy, so your product and your agent share one connection per user.
Plug DeepWiki into your agent.
Your users connect DeepWiki once, under your brand. Your agent gets 3 actions behind your policies, with every call on the record.
Free to build · MCP + REST · Audit log on every plan