dify-canvas-mcp
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
dify-canvas-mcp
MCP server for Dify workflow canvases. Read, patch, publish and test drafts through the user's signed-in Chrome — in any MCP client (Qoder, DSH, Claude Desktop, …).
Tools
Tool | What it does |
| List configured instances |
| Add/update an instance (alias + site URL only) |
| Show which browser channel is available |
| Start a dedicated Chrome with a debugging port |
| List apps (an app can be referenced by name later) |
| Draft node index (+ hash), or one node in full |
| Patch nodes; dry-run by default, |
| Publish and verify the published hash |
| Run the draft once and return the executed node path |
Related MCP server: BrowserLemon
Install
npm i -g dify-canvas-mcp
# or run without installing:
npx -y dify-canvas-mcpMCP client config (stdio):
{
"mcpServers": {
"dify-canvas": {
"command": "npx",
"args": ["-y", "dify-canvas-mcp"]
}
}
}First use
Make sure Chrome is signed in to your Dify site. Preferred: the chrome-pool multiplexer on
127.0.0.1:3333(one authorization, reused) — or your daily Chrome started with--remote-debugging-port=9222.Ask the agent to call
dify_browser_status; if nothing is available, calldify_browser_launchto open a dedicated Chrome window and sign in there once.dify_setup { alias: "prod", baseUrl: "https://dify.example.com" }.From then on, say the app's name — no UUID needed.
Credentials
The credential is the browser session. No cookie or token is stored on disk. Config lives in ~/.dify-canvas/config.json; backups of every written draft go to ~/.dify-canvas/backups/.
Write safety
dify_canvas_patchis dry-run unlessapply:true.Pre-write: on-disk backup + self-check (dangling edges, self loops,
default_valuetypes, code-node parameter mismatch).Post-write: page reload, rendered-node count, crash banners.
dify_canvas_publishverifies the published hash against the draft hash.
MIT.
This server cannot be deployed
Maintenance
Related MCP Connectors
Build and run AI workflows, agents, data tables and 1000+ API integrations from any MCP client.
1Create, browse, remix, collaborate on, and run durable AI workflow nodes from MCP hosts.
Browser MCP for logged-in tasks. Uses your Chrome — credentials stay local. Zero-token replay.
Zero-setup MCP gateway securely connecting AI to your tools with authentication and workflows
Related MCP Servers
- AlicenseNot gradedqualityCmaintenanceConnects AI agents to your Chrome browser via MCP, enabling real-time control of existing tabs, sessions, and application state for development workflows.MIT
- AlicenseBqualityCmaintenanceEnables AI to control your existing logged-in Chrome session via MCP, supporting social media drafting and explicit user-confirmed publishing.8MIT
- AlicenseNot gradedqualityBmaintenanceEnables remote AI agents to control a local, authenticated desktop browser over MCP through an outbound WebSocket bridge and Chrome extension, supporting tab management, navigation, page reading, clicking, typing, scrolling, screenshots, and JavaScript evaluation.1MIT
- FlicenseNot gradedqualityCmaintenanceEnables a local MCP server to control an authenticated Chrome browser for interacting with Google Search, Gemini, ChatGPT, Claude, and DeepSeek, including web search and image generation without per-provider API keys.-