Dify for AI agents
Dify lets teams run published AI workflow apps through a service API and inspect each run's configuration and history, used by developers embedding chatbots and agents into products. 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 Dify.
Trigger a published workflow
Run a workflow app with new input to generate an output a user requested.
Check a workflow run's result
Look up a specific workflow run to confirm it finished and see what it produced.
Review run history first
List recent workflow runs as a read-only check before deciding whether to trigger another.
4 Dify actions, graded by risk.
Every Dify 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.
- dify.get_workflow_runGet Workflow Run
- dify.list_workflow_runsList Workflow Runs
- dify.get_workflow_app_configurationGet Workflow App Configuration
write
1Create and change things. Allow, or ask the user first.
- dify.run_workflowRun Workflow
destructive
0Delete, cancel or archive. Ask first, or deny outright.
- No destructive actions.
Dify in three steps.
- 01Your users connect DifyThey add their Dify api key on Arc0 Connect, under your brand. It goes straight into the vault.
- 02You set the rulesReads run, and writes like “run Workflow” can wait for the user to approve.
- 03Any agent can actYour agent calls Dify through the Arc0 SDK or MCP, and so can Claude, ChatGPT and Cursor. Every call lands on the audit log.
await arc0.policies.set('dify', { read: 'allow', write: 'ask', // run_workflow destructive: 'deny', }) # Claude Code: the same connection, one URL $ claude mcp add --transport http arc0 \ https://mcp.arc0.ai/u/u_8f2
How Dify connects.
Users add their Dify 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 Dify 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 Dify from any agent.
ClaudeChatGPTCursor
Codex
VS CodeOpenAI Agents SDKClaude Agent SDK
Vercel AI SDK
Mastra
LangGraph
Dify and Arc0, answered.
Can I use Dify with Claude, ChatGPT or Cursor?
Yes. Connect Dify 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 Dify actions you allow.
How do users connect Dify?
Users add their Dify 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 Dify actions can my agent take?
4 in total: 3 read, 1 write and 0 destructive, such as “run Workflow”. Your policies decide which of them each agent may call.
Can I make my agent read-only in Dify?
Yes. Allow read actions and deny writes in the Dify policy. Your agent can still look things up, and any write it attempts is blocked and logged.
Can my own backend call Dify too?
Yes. The same Dify connection is available over REST and through the Arc0 proxy, so your product and your agent share one connection per user.
Plug Dify into your agent.
Your users connect Dify once, under your brand. Your agent gets 4 actions behind your policies, with every call on the record.
Free to build · MCP + REST · Audit log on every plan