MLB and NBA, FanDuel and DraftKings — projections, ownership, configured simulations, and availability/blowout checks built in.
Build DFS lineups by talking to an AI agent
Every DFS optimizer until now is a web app full of sliders you set by hand. V12 flips it: describe the lineups you want in plain English and an AI agent builds them — or wire V12 into Claude, Cursor, or any agent via MCP and generate contest-ready FanDuel and DraftKings lineups from your own workflow.
The manual optimizer vs the AI way
A traditional optimizer makes you open the tool, set min and max exposure, lock and ban players, define stack rules, pick a sim count, and export — every slate, by hand. It works, but it’s slow, and the settings are where mistakes creep in.
With V12 you say what you want and the agent handles the knobs. “Build 20 GPP lineups stacking the top run environments, max 4 per team, fade the chalk shortstop” — V12 reads the slate, applies the constraints, runs the sims, and hands back an upload-ready CSV. That’s the difference between a calculator and an assistant that knows the slate.
What you can ask V12
Drive it from your own AI agent
V12 ships an MCP server— the open standard AI agents use to call tools. Connect it to Claude, Cursor, or any MCP client and your agent can read the slate, build candidate pools, run configured simulations, pull ownership and leverage, and export the FanDuel CSV, all as tool calls. It’s the only DFS optimizer you can hand to an agent instead of clicking by hand. For builders, there’s also a JSON CLI and an HTTP API.
Why ChatGPT alone can’t do it
A general chatbot has no live slate, no salaries, no projections, and no optimizer — so it confidently invents players and prices that don’t exist, and the “lineup” is often over the cap or includes someone who isn’t playing. The model is good at language, not at solving a salary-capped optimization over tonight’s real player pool.
V12 closes that gap: it gives the agent a real engine and live data to call. You get the natural-language interface of a chatbot with the correctness of a true optimizer — valid, contest-ready lineups built from today’s actual board.
What’s under the hood
Ask for the lineups you want; the agent translates it into exposure caps, stacks, locks, and constraints for you.
Drive V12 from Claude, Cursor, or any AI agent over the Model Context Protocol — or your own scripts via the CLI and HTTP API.
Every build exports a FanDuel/DraftKings entry template with player IDs and entry IDs — one click to the contest.
AI DFS, answered
Can AI actually build DFS lineups?
Yes — but only when the AI is connected to a real optimization engine with live data. V12 is that engine: it pulls the current slate, projections, ownership, and salaries, runs the optimizer, and returns valid, contest-ready lineups. The AI agent is the interface; the math is real.
Is this just ChatGPT for DFS?
No. A general chatbot has no live slate, no salaries, no projections, and no optimizer, so it will invent players and prices that don't exist. V12 gives the agent a real tool to call — so the lineups are built from today's actual data, not guessed.
Do I need to know how to code?
No. Open the dashboard and type what you want. The MCP server and API are there if you DO want to wire V12 into Claude, Cursor, or your own tools — but they're optional.
What is an MCP server?
The Model Context Protocol is an open standard that lets AI agents call external tools. V12 ships an MCP server, so any MCP-compatible agent can generate lineups, read the slate, and pull ownership/projections as tool calls.
Which sports and sites?
MLB and NBA, on FanDuel and DraftKings. NFL is the most-requested addition.
Try the AI-agent optimizer
MLB and NBA, FanDuel and DraftKings. 7-day free trial, then $99/mo.