Input
Your games
PGNs, URLs, and live games enter the same prep pipeline.
Match-day preparation room
Turn real games into a match-day plan you can trust.
Import your own games, clean the corpus, model the opponent, validate one practical novelty, rehearse the critical positions, and leave with a pack you can review before the round.
Local analysis by default. AI coaching stays behind first-party server boundaries.
Prep Chain
Evidence
Output
Match-Day Pack
Focus lines, drills, directives, provenance.
Input
Your games
PGNs, URLs, and live games enter the same prep pipeline.
Evidence
Trust bands
Engine disagreement and corpus quality stay visible.
Output
Pack
The end state is an exportable match-day artifact.
Preparation chain
The experience should feel like an evidence desk: every action moves the player closer to a concrete match-day pack.
Bring PGNs, public game URLs, or live games into one local-first corpus.
Source material
Expose repertoire branches, move-order tendencies, and corpus confidence.
Opponent map
Rank candidate deviations by surprise, practicality, and eval tolerance.
Prepared choice
Convert swing positions into drills and readiness checks.
Execution reps
Assemble focus lines, directives, checklist, and provenance into one portable artifact.
Match-day pack
Product surface
ChessLM should feel like a tournament scoresheet merged with an engine lab: compact, legible, and built around decisions instead of decoration.
Corpus, model, novelty, critical moments, drills, readiness, and pack stay visible as the working loop.
Analysis depth, engine disagreement, and reliability warnings use explicit labels plus semantic color.
Text and voice coaching must cite the current board, corpus, or prep artifact rather than free-floating advice.
React Three Fiber keeps the premium board native to the workspace without a separate game runtime.
Capture gate
The launch page should show product media only after the capture is tied to a real workflow, redacted, and approved. Until then, the page uses a data-neutral product composition instead of demo screenshots.
Demo-only captures are deliberately excluded from this page. Add approved, redacted production captures before any award or launch review.
Capture corpus quality, novelty validation, readiness, pack export, and coach grounding from production flows.
Account IDs, prompts, transcripts, tokens, and private game data must be removed before publishing.
Development preview data can exist inside the app, but cannot be used as judged marketing proof.
Release bar
Award-level polish fails if storage, auth, engine, payment, or provider boundaries degrade silently.
Provider calls stay behind first-party server or worker routes. No keys ship in browser bundles.
Storage, auth, engine, API, and payment failures must surface visible states with machine-classifiable codes.
Canvas health, board readability, responsive layout, auth, payment, and build gates decide launch readiness.
Pricing
The workspace is free. You only pay for AI usage (chat + voice).
Monthly pricing
$0/mo
Full local prep workspace. No included AI credits.
$8/mo
Includes AI credits for regular training and prep.
$20/mo
For heavy prep cycles and longer voice sessions.
Bring games or a position. Leave with evidence, drills, and a match-day pack.