planeshift-mcp
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@planeshift-mcp add a note to the Apollo project: client approved the new scope"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
planeshift-mcp
MCP server for Planeshift. It runs locally over stdio and forwards tool calls to the Planeshift Rails API (/api/v1).
Install
Requires Node.js 20 or newer.
1. Log in (once)
npx -y planeshift-mcp loginYour browser opens Planeshift. Sign in, click Allow, and the token is saved to ~/.config/planeshift/credentials.json (readable only by you). Claude Code and Claude Desktop both read it, so you only do this once. Use --url https://your-instance for a self-hosted Planeshift. npx -y planeshift-mcp logout revokes the token and deletes the file.
2a. Claude Code
claude mcp add --scope user planeshift -- npx -y planeshift-mcpStart a new claude session and run /mcp to check that planeshift is connected.
2b. Claude Desktop
Open Settings > Developer > Edit Config (or edit ~/Library/Application Support/Claude/claude_desktop_config.json on macOS, %APPDATA%\Claude\claude_desktop_config.json on Windows) and add the server:
{
"mcpServers": {
"planeshift": {
"command": "npx",
"args": ["-y", "planeshift-mcp"]
}
}
}Quit Claude Desktop completely and open it again. The tools appear under the tools icon of the chat box. If npx is not found, use the full path of npx (run which npx) as the command, since the app does not load your shell profile.
Tools
list_projects, get_timeline, add_note, add_meeting, add_todo_action, add_important_link.
Development mode
To use a local Planeshift (http://localhost:3000) without touching your production login, set PLANESHIFT_ENV=development, or pass --dev to the commands:
node src/index.js login --dev # saves ~/.config/planeshift/credentials.development.json
claude mcp add planeshift -e PLANESHIFT_ENV=development -- node /path/to/planeshift-mcp/src/index.jsFor Claude Desktop, use "command": "node" with the full path to src/index.js and "env": { "PLANESHIFT_ENV": "development" }. Development mode keeps its own credentials file, and --url still overrides the address.
Without login
Set PLANESHIFT_URL and PLANESHIFT_TOKEN (create a token on the API tokens page) in the env of the config above, or with claude mcp add -e PLANESHIFT_URL=... -e PLANESHIFT_TOKEN=.... They take precedence over the saved credentials.
Related MCP server: Nexxupp MCP Server
Development
npm install
npm startAdd a tool in src/tools/ and register it in src/index.ts.
Available Tools
6 toolsadd_important_linkB
Add a reference (an important link, like a dashboard or a doc) to a project.
Inputs: url (required), name (required), project_id (required).
add_meetingB
Record a meeting on the timeline of a project: what was discussed and decided.
Inputs: project_id (required), description (required), date, title.
add_noteC
Add a note to the timeline of a project.
Inputs: project_id (required), description (required).
add_todo_actionB
Add an action (a to-do item) to a project. It starts as pending.
Inputs: title (required), project_id (required), description.
get_timelineA
Read the timeline of a project (notes, meetings and actions), newest first. Pass query to search titles and descriptions.
Inputs: project_id (required), type, limit, query.
list_projectsA
List the Planeshift projects of the current user. Use it to find a project id.
Inputs: no named parameters.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 6 tools
The tools split cleanly into two reads (list_projects, get_timeline) and four distinct add_* writers targeting different artifacts (note, meeting, action, link). The only mild overlap is add_note vs add_meeting, since both append to the same timeline, but the descriptions clarify that meetings capture discussion/decisions.
All names are snake_case verb_noun (list_projects, get_timeline, add_note...), which is predictable and readable. Minor deviation: add_todo_action and add_important_link insert extra modifiers, making the noun phrase less uniform than the others.
Six tools is well-scoped for a project timeline server: two read operations plus four creation operations covering the domain's core entities. Every tool earns its place with no filler.
The surface covers listing, reading, and appending, but there is no way to update or delete notes/meetings/links, and notably no way to mark a to-do action as completed despite actions starting as pending. These gaps will force agents into dead ends for a to-do workflow.
Maintenance
Related MCP Connectors
Persistent context for Claude. Your AI always knows your projects and next actions across sessions.
Connect Claude to your Platform7n workspaces: chat, links, work and decks. One-click OAuth.
Connect Claude to your Platform7n workspaces: chat, links, work and decks. One-click OAuth.
Read your tracked time and projects, pause tracking or add a note, from Claude or ChatGPT.
Related MCP Servers
- FlicenseCqualityDmaintenanceConnects Claude to a local project management API, enabling task, project, and user management through natural language.46-
- AlicenseBqualityCmaintenanceEnables AI agents to store and search private memory, explore and edit workspace stations, databases, tasks, and charts, and schedule calendar events through natural language. Supports local stdio and hosted HTTP modes with per-user OAuth scoping of tools.22MIT
- AlicenseNot gradedqualityAmaintenanceEnables AI clients such as Claude, Cursor, VS Code, and Windsurf to manage workspaces, projects, tasks, and AI agents through hosted OAuth or local stdio tools.2,530 npmMIT
- AlicenseNot gradedqualityAmaintenanceConnects a workspace and project management platform to Claude, Cursor, VS Code, Windsurf, and other MCP clients, letting users list workspaces, create and manage projects and tasks, and build, train, and prompt AI agents. Also generates MCP tools from any OpenAPI spec.2,530 npmMIT