About niftypm
Nifty's hosted MCP server connects AI agents to workspace data and actions, including tasks, projects, conversations, documents, and reporting. Its domain-based tools cover searching, changes, and deletion, with access limited by the signed-in user's Nifty permissions.
Key features
- 01Search workspace information
- 02Create and update workspace data
- 03Delete workspace records
- 04Access tasks, projects, and messages
- 05Manage documents, people, and reports
Use cases
- 01Find overdue tasks by project
- 02Create tasks from meeting notes
- 03Summarize project status and risks
- 04Search workspace documents and files
MCP support
Exposes Nifty workspace operations as MCP tools over hosted Streamable HTTP. Clients connect with a personal access token or Nifty's account connector; calls follow the user's Nifty permissions.
Configuration
Remote servers need no local install. Pick your client, then copy the config or install in one click.
https://mcp.niftypm.com/mcpclaude mcp add --transport http com-niftypm https://mcp.niftypm.com/mcp{"mcpServers":{"com-niftypm":{"type":"http","url":"https://mcp.niftypm.com/mcp"}}}Open in Cursor{"servers":{"com-niftypm":{"type":"http","url":"https://mcp.niftypm.com/mcp"}}}Open in VS Code{"mcpServers":{"com-niftypm":{"type":"http","url":"https://mcp.niftypm.com/mcp"}}}{"mcpServers":{"com-niftypm":{"serverUrl":"https://mcp.niftypm.com/mcp"}}}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.
https://mcp.niftypm.com/mcpAuth requiredTools
Details
From the project READMENifty MCP Server
The official Model Context Protocol server for Nifty. It gives AI agents access to your Nifty workspace: tasks, projects, messages, documents, files and more. It works with Claude, ChatGPT, Cursor and any other client that supports remote MCP servers.
This is a hosted (remote) server. There is nothing to install or run. This repository holds its
public description, registry metadata (server.json) and connection guide. The server itself is
operated by Nifty.
Connect
Endpoint: https://mcp.niftypm.com/mcp (Streamable HTTP)
Easiest: in Nifty, open Settings → MCP (app.nifty.pm/settings/mcp). The wizard covers
Claude, Claude Code, ChatGPT and Cursor, with a ready-to-paste config or an OAuth connector flow.
Manually: add the endpoint as a remote MCP server and send a Nifty Personal Access Token as a
Bearer token. Create one at app.nifty.pm/settings/personal-access-tokens:
Authorization: Bearer <your Nifty personal access token>
Claude, ChatGPT and Microsoft 365 Copilot can instead sign in with your Nifty account through Nifty's connector, started from Settings → MCP. Other clients need the token.
The plugin and extension manifests in this repository read the token from the NIFTY_PAT
environment variable (Gemini CLI asks for it when you install the extension):
export NIFTY_PAT=<your Nifty personal access token>
Claude Code example:
claude mcp add --transport http nifty https://mcp.niftypm.com/mcp --header "Authorization: Bearer $NIFTY_PAT"
Install as a plugin or extension
This repository is also a plugin for the clients that install from a Git repo:
| Client | Install |
|---|---|
| Claude Code | /plugin marketplace add Nifty-Technologies-Inc/nifty-mcp, then /plugin install nifty@nifty |
| Gemini CLI | gemini extensions install https://github.com/Nifty-Technologies-Inc/nifty-mcp |
| Cursor | Copy mcp.json into .cursor/mcp.json, or add the endpoint in Settings → MCP |
| Cline | See llms-install.md |
Clients that only support stdio
Bridge with mcp-remote. Keep the space out of args by
putting the whole header value in env:
{
"mcpServers": {
"nifty": {
"command": "npx",
"args": ["-y", "mcp-remote", "https://mcp.niftypm.com/mcp", "--header", "Authorization:${NIFTY_AUTH}"],
"env": { "NIFTY_AUTH": "Bearer <your Nifty personal access token>" }
}
}
}
Example prompts
- "What's overdue in the Website Redesign project? Group it by assignee."
- "Create a task in Marketing called 'Draft Q4 launch post', assign it to me, due Friday."
- "Summarize this week's messages in the Product discussion."
- "Move every task tagged 'bug' in the Mobile App project to the In Review status."
- "Which milestones in the Client Onboarding project are at risk?"
- "Find the doc about our release checklist and list its open items."
Tools
Tools are grouped by domain, and each domain has three verbs: _query (browse, filter, paginate,
full-text search), _mutate (create, update, link relations, bulk writes) and _destroy (delete).
| Tool | What it does |
|---|---|
tasks_query | Find tasks, checklists and reminders |
tasks_mutate | Create or update tasks, checklists and reminders |
tasks_destroy | Delete tasks, checklists and reminders |
planning_query | Find projects, portfolios, milestones and goals |
planning_mutate | Create or update projects, portfolios, milestones and goals |
planning_destroy | Delete projects, portfolios, milestones and goals |
communication_query | Find chats, messages and check-ins |
communication_mutate | Create or update chats, messages and check-ins |
communication_destroy | Delete chats, messages and check-ins |
knowledge_query | Find docs and files |
knowledge_mutate | Create or update docs and files |
knowledge_destroy | Delete docs and files |
people_query | Find members |
people_mutate | Invite or update members |
people_destroy | Remove members |
reporting_query | Find time entries, charts and project activity |
reporting_mutate | Create or update time entries and charts |
reporting_destroy | Delete time entries and charts |
admin_query | Find labels, roles, favorites and custom fields |
admin_mutate | Create or update labels, roles and favorites |
admin_destroy | Delete labels, roles and favorites |
The full machine-readable list (descriptions and input schemas) is published at
mcp.niftypm.com/.well-known/mcp/server-card.json.
Every call runs as the signed-in user (or the token's agent identity), with that user's Nifty permissions. The server only sees what you can see.
Prompts
plan_my_week: Draft a realistic day-by-day plan for the week from your open, due-soon tasks.project_status_digest: Summarize a project’s status: progress, overdue items, recent shipments, and top risks.triage_overdue_tasks: Work through your own overdue tasks and decide what to reschedule, reassign, drop, or do now.standup_digest: Draft a Yesterday / Today / Blockers standup update from your tasks.meeting_notes_to_tasks: Turn meeting notes into well-formed tasks in a project, after your review.find_and_update: Find the tasks matching a description and apply a change — with a review step first.
Docs and support
- Documentation: Nifty MCP Server guide
- Support: nifty.pm
License
The contents of this repository are MIT-licensed (see LICENSE). The hosted service is covered by Nifty's terms of service.
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.