Skip to main content
Glama

Edworking MCP

Edworking MCP on Glama

The official, open-source Edworking MCP server. Connect Claude, ChatGPT, Cursor and other AI assistants to the projects, tasks, Docs, conversations and files your Edworking account can access. Your assistant answers from your live workspace, with links back to Edworking, and takes the next step when you ask.

What you can do

  • Find the work: browse projects, statuses, members and filtered tasks. Search indexed tasks, Docs, files and project messages, and open a result with its link.

  • Move tasks forward: create projects and tasks, update a task's name, status, priority, assignees, start date and deadline, and make targeted edits to task descriptions.

  • Write Docs: create, read and rename Docs, and make targeted Markdown edits.

  • Follow up in conversations: read accessible conversations and send the message you ask for to a specific one.

  • Files and designs: list file details, import a public HTTPS file, add Figma designs to new tasks and Docs, link a frame to a task, and find the tasks linked to a Figma file.

Try asking:

  • "What's overdue in the Launch project, and who owns each task?"

  • "Find where we discussed onboarding pricing and summarize it with links."

  • "Create tasks in Launch from these meeting notes and set Friday deadlines."

  • "Read the latest messages in the planning conversation and summarize the decisions."

Deleting projects, tasks or Docs, inviting people and changing sharing are not available.

Related MCP server: Todocko MCP Server

Connect

Add https://mcp.edworking.com/mcp to your assistant as a remote MCP server and start its OAuth connection. On the authorization page, choose Continue in Edworking and sign in. Pick your workspace, review the assistant, its destination and the permissions, then choose Allow access. Edworking creates a scoped token for the connection, named after your assistant, that expires within 30 days.

If you can't use Edworking sign-in, create a dedicated token in Edworking → Profile → API tokens and enter it only on the authorization page. Never paste a token into an assistant conversation or client configuration.

Assistant

Setup

Claude

Settings → Connectors, add a custom connector with the server URL. Choose Sign in now and Use Claude's published identity, and leave request headers empty. No client secret is needed.

Claude Code

claude mcp add --transport http edworking https://mcp.edworking.com/mcp, then run /mcp to authorize.

ChatGPT

Plugins → Add custom MCP server, with OAuth and the server URL. Availability depends on your plan and workspace policy.

Codex

codex mcp add edworking --url https://mcp.edworking.com/mcp, then codex mcp login edworking.

Cursor, VS Code, Windsurf and others

Add a remote HTTP server named edworking with the server URL. The setup guide has each client's configuration.

Clients authenticate as public OAuth clients with PKCE S256, using dynamic registration or an HTTPS Client ID Metadata Document. For Claude's published identity, Edworking negotiates only the authorization-code and refresh-token grants. /mcp accepts only Edworking MCP access tokens.

Permissions

New connections request all 11 supported read and write permissions by default, enabling the full 24-tool catalog after approval. The authorization screen lists each permission before you connect. Writes still require the matching OAuth scope and upstream token permission. Clients can request a smaller scope set, including read-only access; explicit requests are never expanded. Existing read-only connections must disconnect and reconnect to approve writes. If a client retains an explicit read-only scope list, update that list to include the required write scopes before reconnecting. Refreshing a token does not add permissions. Revoke the connection's token in Edworking to disconnect every grant using it. Grants expire within 30 days or when the Edworking token expires, whichever is sooner. Access tokens last up to ten minutes and refresh tokens rotate.

Troubleshooting

After Allow access or Cancel, the browser returns to your assistant automatically. If it remains on Returning to your assistant, use the return button. Each consent request can be used once and expires after ten minutes. If you see invalid_grant or an expired/used request, start a new connection from your assistant instead of reopening the old consent link. Do not refresh or resubmit the approval form.

Tools

Tool

Scope

Purpose

get_connection

Any connection

Connection name, permissions and expiry

list_projects, list_project_statuses, list_project_members

projects:read

Projects, valid statuses and existing assignee IDs

list_tasks, get_task

tasks:read

Filter tasks by project, status, assignee, priority, title and due dates; read task content

list_figma_tasks

tasks:read

Find accessible tasks linked to a Figma file or frame

link_figma_frame

tasks:write

Link a Figma frame to an accessible task

search, fetch

Per-source read scope

Indexed tasks, Docs, files and project messages; typed result retrieval

list_docs, get_doc

docs:read

Find Docs and read their Markdown content

create_doc, update_doc, edit_doc_content

docs:write; body edits also need docs:read

Create, rename and edit Docs

list_conversations

chats:read

Existing private conversations

read_messages

messages:read

Accessible conversation messages

list_files

files:read

File metadata

create_task, update_task, edit_task_content

tasks:write; body edits also need tasks:read

Create tasks; edit name, status, priority, assignees, dates and description

send_message

messages:write

Send the user's specified message

create_project

projects:write

Create a project

upload_file

files:write

Import a public HTTPS file

Attaching files to tasks also requires files:write. Lists use cursors and at most 100 items per page; keep ordering and filters unchanged while paging. Search uses the workspace index and returns up to 20 ranked results per permitted source. It rechecks live access, deletion state and message-history limits before exposing snippets. Indexing may lag; unavailable sources are reported, and private chats and projectless items are excluded. Use filtered lists for exhaustive task discovery. File fetches return metadata, not file contents.

Read a task or Doc before editing its body. Content reads are capped at 20,000 characters and expose block handles; targeted edits reuse the collaboration service and version history. Large removals require explicit allowRemoval. A task assignee list replaces the current list: omit it to preserve, or use [] to unassign. Assignees must already belong to the project and workspace. Deletion, invitations, task moves and sharing changes are not exposed.

Writes require a UUID requestId; reuse it with unchanged arguments on retries. Completed results are encrypted and cached for 24 hours. An uncertain result is never repeated automatically: inspect Edworking before deciding what to do. This is not an upstream exactly-once transaction. Deduplication expires after 24 hours.

Development

Use Node 24 LTS (minimum 22.19). No OpenAI API key is needed.

npm ci
cp .env.example .env
# Generate a key and save it as TOKEN_ENCRYPTION_KEY in .env:
node -e "console.log(require('crypto').randomBytes(32).toString('hex'))"
npm run build
node --env-file=.env dist/index.js

Run npm run check for type checking, security/integration tests and build. npm run dev starts a watcher when variables are already loaded by your shell.

For local stdio, configure your client to run npx -y @edworking/mcp with EDWORKING_API_TOKEN supplied securely in its process environment, or build from source and run node /absolute/path/to/edworking-mcp/dist/stdio.js. The npm package is public. Stdio uses that token directly with Edworking and has in-memory deduplication that resets on restart.

Railway deployment

Use the included Dockerfile, attach a volume at /data, and set:

Variable

Value

NODE_ENV

production

PORT

3000

PUBLIC_BASE_URL

Stable public HTTPS origin, without /mcp

DATABASE_PATH

/data/bellsprout.sqlite

TOKEN_ENCRYPTION_KEY

Random 32-byte key encoded as 64 hex characters

EDWORKING_API_URL

https://gateway.edworking.com/

EDWORKING_APP_URL

https://app.edworking.com

Keep one replica, sleeping off and deployment overlap zero. /healthz checks the process and /readyz checks SQLite. Preserve the key across deployments. Optional ALLOWED_ORIGINS adds exact browser origins, OPENAI_DOMAIN_CHALLENGE serves the OpenAI verification value, and EDWORKING_CONNECT_URL enables a native Edworking connection page only after that page has shipped.

Figma

The Edworking Figma plugin and Figma's AI connectors also use this server. The plugin transport, native consent, the figmaDesigns input, retry behavior and deployment order are in docs/figma.md. The server page has Figma / Figma Make connector steps.

More

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      Not graded
      quality
      B
      maintenance
      Enables AI assistants to manage Todocko application data, including tasks, projects, worklogs, and attachments. It supports comprehensive project management operations such as tracking activity, managing Kanban boards, and handling shared project synchronization.
      253 npm
      1
      MIT
    • A
      license
      A
      quality
      D
      maintenance
      Enables AI assistants to interact with the Worksection project management platform to manage projects, tasks, comments, and members through natural language. It provides a comprehensive suite of tools for task execution and project oversight while automatically handling API rate limits.
      21
      MIT