AI review for game design docs
How Gameframe surfaces inconsistencies, contradictions, and missing detail in a design doc using a model run that respects the source-of-truth contract.
Gameframe runs a model pass on every new version of a design doc. The output is not a rewrite. It is a structured diff against the previous version plus a short list of flagged inconsistencies the model thinks the human reviewer should look at before merging.
Three things make the AI pass useful instead of noisy:
- Results are cached by content hash. Re-running the same doc costs nothing within the cache TTL.
- The prompt carries the full version history of the doc — not just the latest diff — so the model can spot a regression against an earlier decision the team thought was settled.
- Credits are reserved before the call is dispatched. A failed upstream call refunds the held credit in full.
AI review is a tool. The reviewer still merges.