Povver MCP Server
Connect your AI assistant to your Povver strength-training data — Claude, Claude Code, Cursor, or any MCP-compatible client. Read workout history, track strength progress, analyze muscle-group volume, and manage routines and templates from a conversation.
Endpoint: https://mcp.povver.ai/mcp (Streamable HTTP, OAuth 2.1)
Docs: https://mcp.povver.ai/ · /llms.txt · /llms-full.txt
Requires: a Povver account with a premium subscription (iOS app)
This is a hosted, remote server — there is nothing to install. Sign in with your Povver account when your client prompts you.
Povver also publishes strength calculators: RPE and RIR, one rep max, DOTS, FFMI, strength standards and weekly sets, each with its method and sources.
Listed on: Official MCP Registry (ai.povver/mcp-server) · Glama · Smithery · awesome-remote-mcp-servers
Connect
Claude (Desktop / web)
Settings → Connectors → Add custom connector → URL https://mcp.povver.ai/mcp. Claude opens a browser window to sign in with your Povver account.
Claude Code
claude mcp add --transport http povver https://mcp.povver.ai/mcp
Cursor / other clients
{
"mcpServers": {
"povver": {
"url": "https://mcp.povver.ai/mcp"
}
}
}
API key (automation, scripts, n8n)
In the Povver app: More → Integrations → Developer access → create a key, then send it as a Bearer token:
curl -H "Authorization: Bearer pvk_your_key_here" \
-H "Content-Type: application/json" \
-H "Accept: application/json, text/event-stream" \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}' \
https://mcp.povver.ai/mcp
Example prompts
- "Act as my strength coach: review my recent training, find my biggest weakness, and update my program to fix it."
- "How's my bench press trending over the last 8 weeks — am I still progressing or have I plateaued?"
- "Break down my weekly training volume by muscle group. Where am I under-training?"
- "What should I train today?"
Available tools
One tool per kind of thing (your context, plans, workouts, the session in progress, analysis, exercises, memory, periodization). Each takes a view (reads) or an action (writes) and only that choice's fields. Every tool carries a display title and readOnlyHint / destructiveHint annotations, and delete is the only destructive one.
Changed in 2.0.0: the single-purpose tools of 1.x were consolidated into these (get_workout is now read_workouts with view: detail, create_template is save_plan with action: create_template, and so on), and explain_rule became the resource povver://rules/{rule_id}. The old names are not kept as aliases: a saved prompt that names one needs the new name. Clients pick up the new list when they reconnect.
Read (safe to call freely)
Write (change your data)
Resources
Attachable as context from a client's "+" menu:
- Training Status — weekly adherence, next workout, days since training
- Active Routine — current program structure
- Recent Workouts — last 10 session summaries
- Training Analysis — AI insights and weekly review
- Individual workouts and templates — browsable by ID
- Tool guide (
povver://guide/tools) and Periodization guide (povver://guide/periodization) — how to import history, edit templates without erasing them, run a live session, read volume - Training rules (
povver://rules/{rule_id}) — the evidence behind each of Povver's training rules, the same cards the in-app coach answers from
Prompts
- Analyze my training week — volume, adherence, and recommendations
- What should I train today? — next workout recommendation
- Review exercise progress — strength trends for a specific lift
- Review my training program — full program structure evaluation
Security & privacy
- OAuth 2.1 with PKCE and dynamic client registration; API keys are hashed at rest and revocable from the app.
- Every request is scoped to the signed-in user; tools only ever touch that user's data.
- Privacy policy: https://povver.ai/privacy · Terms: https://povver.ai/tos
- Support: mailto:[email protected]
The server source is part of Povver's private codebase; this repository holds the public documentation and the MCP Registry manifest (server.json).