About Tallyfy Workflow Automation
Tallyfy MCP Server is the official MCP server for Tallyfy's workflow platform, enabling AI assistants to work with tasks, processes, templates, and organization data. It uses the user's Tallyfy OAuth identity and limits access to that user's permissions.
Key features
- 01Create, update, and search tasks
- 02Launch and track processes
- 03Read and edit templates and steps
- 04Manage people, groups, and organization data
- 05Read and upload attached files
Use cases
- 01Advance process steps through an AI assistant
- 02Find and reassign tasks
- 03Edit templates and fill form fields
- 04Search organization data or product documentation
MCP support
This server connects MCP clients to Tallyfy; its hosted endpoint uses streamable HTTP and Tallyfy OAuth. Tool access is limited by the signed-in user's Tallyfy permissions.
Configuration
Remote servers need no local install. Pick your client, then copy the config or install in one click.
https://mcp.tallyfy.com/claude mcp add --transport http com-tallyfy https://mcp.tallyfy.com/{"mcpServers":{"com-tallyfy":{"type":"http","url":"https://mcp.tallyfy.com/"}}}Open in Cursor{"servers":{"com-tallyfy":{"type":"http","url":"https://mcp.tallyfy.com/"}}}Open in VS Code{"mcpServers":{"com-tallyfy":{"type":"http","url":"https://mcp.tallyfy.com/"}}}{"mcpServers":{"com-tallyfy":{"serverUrl":"https://mcp.tallyfy.com/"}}}Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.
Tools
Details
From the project READMETallyfy MCP Server
The official Model Context Protocol server for Tallyfy, the workflow and process automation platform. It lets any MCP-capable AI assistant run your operations: create and track tasks, launch and advance processes, read and edit templates, fill form fields, manage people and groups, and search across your organization, all through your own authenticated Tallyfy account.
This repository is an automatically published, read-only mirror. The canonical source is maintained privately by Tallyfy and a sanitized, server-only snapshot is published here on every production release. Please do not open pull requests against this mirror (they cannot be merged here). For bugs or feature requests, use the issue tracker or contact us (see Support). Hosting, deployment, and monitoring code is intentionally not part of this mirror.
Connect (hosted, recommended)
Tallyfy runs a managed remote server. You do not need to host anything. Point your MCP client at:
https://mcp.tallyfy.com/
Transport is streamable HTTP and authentication is OAuth against your Tallyfy account, so the assistant only ever sees data the signed-in user is allowed to see.
Example client config
{
"mcpServers": {
"tallyfy": {
"type": "streamable-http",
"url": "https://mcp.tallyfy.com/"
}
}
}
What it can do
129 tools across the Tallyfy domain, grouped by area:
| Area | Examples |
|---|---|
| Tasks | create, complete, reassign, comment on, and search tasks |
| Processes (runs) | launch a process from a template now or on a schedule, advance and track steps |
| Templates (checklists) | read, create, and edit templates and their steps |
| Form fields | read and populate kick-off and step form fields |
| Automation | inspect and manage automated actions / rules, explain why a step in a running process is shown or hidden, and list the open tasks for a step across the processes still running |
| People & access | users, groups, guests, organization membership |
| Organization | tags, folders, comments, search across the org |
| Files | read and upload files attached to form fields, kickoff fields and task comments |
| API fallback | read from or write to any Tallyfy REST endpoint that has no dedicated tool |
| Product docs | search the public Tallyfy documentation index |
| Org memory | read and update a per-organization memory document held server-side |
Almost every tool calls the public Tallyfy API on behalf of the authenticated user (the exceptions read the public docs index and the server-side org memory store). There are no write operations the signed-in user could not perform in the Tallyfy app directly.
Doing bulk or scripted work? This server is tuned for interactive, conversational use and limits how much one request can do at once. For high-volume or headless automation (launching processes from a CSV, exporting every template to disk, multi-org loops, CI/CD approval gates), reach for the first-party tallyfy CLI instead. It is built for deterministic, scriptable batch operations. Install with brew install tallyfy/tap/tallyfy or grab a binary from the releases page.
Run it yourself
You can build and run the server from this mirror with Docker. No configuration is required to start it:
cd server
docker build -t tallyfy-mcp-server .
docker run --rm -p 9000:9000 tallyfy-mcp-server
curl http://localhost:9000/health # {"status":"healthy"}
The server listens on port 9000 and speaks streamable HTTP at / (and at /mcp). Python 3.13 is required if you prefer to run it without Docker (pip install -r server/requirements.txt, then uvicorn server:app from inside server/).
To configure it, copy .env.example to .env and pass it with --env-file .env. Every value in there is optional; the defaults point at Tallyfy production.
About authentication
Every MCP request needs a Tallyfy-issued OAuth 2.1 bearer token (RS256). The server verifies that signature on every call, so an unauthenticated client gets a 401 challenge and reaches no Tallyfy data. Only /health and the OAuth discovery documents are open.
The verification key is resolved for you. Left alone, the server reads Tallyfy's published JWKS at https://account.tallyfy.com/.well-known/jwks.json and picks up key rotations automatically. Set TALLYFY_PUBLIC_KEY to a PEM public key if you would rather pin it and avoid the network lookup. If neither source resolves, the server still starts, and rejects every token.
Listed on
- Official MCP Registry (
com.tallyfy/mcp-server) - GitHub MCP Registry
- Smithery
- Glama
- PulseMCP
Security
Tools are scoped to the authenticated user, inputs are validated and the outbound API surface is allowlisted. We disclose no static credentials in this repository: all secrets are supplied at runtime via environment variables. If you find a security issue, please email [email protected] rather than opening a public issue.
Support
- Product docs: https://tallyfy.com/products/pro/integrations/mcp-server/
- Help / contact: https://tallyfy.com/contact/
License
Apache License 2.0 © Tallyfy, Inc.
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.