fpl-strategy-mcp
Provides tools for Fantasy Premier League (FPL) strategy, including player search, fixture outlook, tier classification, hit justification, blank/double gameweek planning, team tracking, and price/ownership trends.
Click on "Install 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., "@fpl-strategy-mcpShould I take a -4 hit for Son?"
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.
fpl-strategy-mcp
A custom Fantasy Premier League MCP server, built specifically to operationalize our squad-building framework — not a generic player-lookup tool. Every tool maps to a named section of the strategy documents this project is built on:
Tool | Framework section it implements |
| foundational data access |
| Layer 1.1 — threshold-hit-rate, not raw totals |
| Layer 3.3 — weighted rolling window, not single FDR |
| Build Strategy §2 — Core Anchor / Value Floor / Edge |
| Build Strategy §4.3 — explicit hit-justification test |
| Build Strategy §8 — chip timing |
| tracking our actual live squad, publicly, no login |
| Layers 1.3 + 2.1 — price/EO signal |
No FPL email or password is ever required. Everything reads public, unauthenticated endpoints only — this was a deliberate choice, partly because we don't need anything else, and partly because handing real login credentials to any third-party code (ours included) is a reasonable thing to be cautious about.
What's been verified vs. what to check on first run
Built and tested in a sandboxed environment with no access to the live FPL API (network allowlist restriction) — so testing here meant:
Full unit tests against synthetic data shaped like real FPL API responses (
test_analysis.py) — 18/18 passing.Full integration tests through the actual MCP tool-call path with a mocked API layer (
test_server_integration.py) — 11/11 passing.Syntax and import verification for every file.
The one thing that's genuinely unverified against the real API is the exact field names for defensive-contribution stats — that's the newest, least publicly documented part of the FPL API. Run this once, after setup, before trusting the DEFCON numbers for real decisions:
python check_live_api.pyIt tells you plainly whether the field-name guesses in analysis.py matched
reality, and exactly what to edit if they didn't. If a field name is wrong, the
tool fails loudly with the actual available keys listed — it will never silently
return a wrong number.
Related MCP server: FPL MCP Server
Setup — Option A: Claude Desktop or Claude Code (recommended, simplest)
This runs entirely on your own machine. Nothing is exposed to the internet, nothing needs hosting, and it's the standard way MCP servers are used locally.
Install Python 3.10+ if you don't have it.
In this folder, install dependencies:
pip install -r requirements.txtRun the live field-name check once:
python check_live_api.pyAdd this to your Claude Desktop config file (macOS:
~/Library/Application Support/Claude/claude_desktop_config.json, Windows:%APPDATA%\Claude\claude_desktop_config.json):{ "mcpServers": { "fpl-strategy": { "command": "python", "args": ["/full/path/to/fpl-strategy-mcp/server.py"] } } }Use the full absolute path to
server.py— relative paths cause silent failures.Restart Claude Desktop completely. You should see the FPL tools available in the tools/hammer icon.
Claude Code works the same way — add the equivalent entry to its MCP config.
Setup — Option B: remote, for use directly inside claude.ai web chat
This chat interface (claude.ai) only connects to MCP servers that are reachable over the public internet — it cannot reach a server sitting on your laptop. To use these tools directly in a claude.ai conversation (rather than Claude Desktop/Code), you'd need to:
Deploy this server somewhere publicly reachable, using HTTP transport instead of stdio. Change the last line of
server.pyto:mcp.run(transport="streamable-http", port=8000)(Small free-tier hosts work fine for this — Render, Railway, Fly.io, etc.)
In claude.ai: Settings → Connectors → Add custom connector, and paste your server's public URL.
This is more setup than Option A for the same result, so start with Option A unless you specifically need it inside this exact chat interface.
Files
server.py— tool registrations (the MCP-facing layer)fpl_client.py— shared API client with caching, no business logicanalysis.py— our custom calculations (DEFCON hit-rate, rolling fixtures, tier classification, hit-math) — this is the part that's genuinely ourscheck_live_api.py— one-time live field-name verificationtest_analysis.py,test_server_integration.py— the test suite; re-run either withpython test_analysis.pyany time you changeanalysis.py
A note on trust
You mentioned not trusting pre-built MCPs for this — that's a reasonable instinct, especially given at least one public FPL MCP server asks for your real FPL email and password to unlock team-viewing features. This server never asks for that, and every file is short enough to read end to end. That's the actual point of building our own: not that public ones are malicious, but that you shouldn't have to take it on faith.
This server cannot be installed
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Related MCP Servers
- AlicenseAqualityCmaintenanceA Model Context Protocol server that provides access to Fantasy Premier League data, allowing users to compare players, find team information, view gameweek data, and get FPL-related advice through Claude for Desktop and other MCP-compatible clients.Last updated1672MIT
- AlicenseBqualityDmaintenanceEnables interaction with Fantasy Premier League data through natural language queries. Supports retrieving fixtures, league standings, player performance, and other FPL statistics via the official FPL API.Last updated16433MIT
- Alicense-qualityDmaintenanceConnects LLMs to the Fantasy Premier League API for intelligent team management, enabling natural language player research, competitor analysis, transfer decisions, and strategic planning using friendly names instead of IDs.Last updatedMIT
- AlicenseAqualityCmaintenanceAI-powered Fantasy Premier League assistant — scored captain picks, transfer suggestions, differentials, fixture outlook, price predictions, live points, and a full manager hub that auto-detects your squad, bank balance, and free transfers.Last updated138MIT
Related MCP Connectors
Read-only fantasy analysis for ESPN, Yahoo, and Sleeper leagues via MCP
Quant tools + an NFL fantasy draft assistant for AI agents — Kalshi & Polymarket EV, edge, ADP.
API-Football MCP — comprehensive soccer/football data
Latest Blog Posts
- Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)By Om-Shree-0709 on .mcpAgent IdentityOAuth 2.1
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/CElyousfi/fpl-strategy-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server