Chess
Counterplay
Eight hand-built, checked chess puzzles and an openings coach that explains every move: the part of chess an AI does well.
- Players
- 1
- Length
- 15–30 min
- Rated
- All ages
- Version
- 1.1.0
- Updated
Works with ChatGPT · Claude · Gemini. Free, no signup.
Opens ChatGPT or Claude with the play prompt ready; press send if it hasn't started. Same link on your phone as on your desktop. Another app? Copy the play prompt into it, or, if it can't open links, copy the full game card below. What works where.
No web access, or want it as an installed app?
The game card is the entire game, about 12k characters: every scene and rule, and the answers the AI referees from. Paste it into an assistant that can't open links, without reading it first: it gives the game away. The Skill installs Counterplay into Claude on any plan: download it, then turn on code execution under Settings → Capabilities and add it at Customize → Skills → + → Upload a skill.
Counterplay is a free chess trainer you play inside ChatGPT, Claude, or Gemini. Fixed-answer chess puzzles and an openings coach, not a full-strength opponent. Solo, 15–30 min a run. No signup and nothing to install: open it in your assistant, paste the play prompt into another, or paste the whole card where links don't open.
How to start
- Open it in your AI. Tap Open in ChatGPT or Open in Claude, and press send, or paste the play prompt into any assistant that can open links. No paid plan required.
- The game loads. Puzzle 1 of eight is on the board as it loads; say "openings" to switch to the openings coach.
- Solve or drill. Work the position or the opening line. The coach explains ideas and nudges you towards the plan.
- Check the answer. Solve it, or give up, and the coach shows the written solution and the one-line reason it works.
Questions
- Can ChatGPT actually play chess?
- Not reliably. Language models predict text rather than legal chess moves, so in a long game they drift into illegal moves and blunders, and in a 2026 benchmark that handed them the full board every turn, even the strongest models still had to be asked to retry illegal or malformed moves. They are useful for coaching, explaining ideas, and drilling puzzles, which is what Counterplay uses them for.
- Why puzzles and openings instead of a full game?
- Because those are the parts an AI can do reliably. Every puzzle's answer is written on the card, so the coach checks your move against a fixed, known solution, and opening lines are known theory the AI can explain. You get accurate positions and a coach, not a shaky opponent.
- How do you know the puzzle answers are right?
- Every puzzle was hand-built, and every solution has been checked against the Stockfish chess engine: each is the engine's best move or a forced mate. The solution ships written on the card, and the coach reads that fixed answer and grades your move against a checked line; it never computes chess itself.
- Do I need to know chess notation?
- No. You can describe moves in plain language, such as "bishop takes c6", and the AI will translate; standard notation such as Bxc6 works too.