Three years building an AI RPG taught this dev to test what changed, not what AI said
tschilpi · reddit · 2026-11-03
Developer tschilpi shares lessons from three years building Lore & Legends Maker, an RPG with an AI Dungeon Master, using Claude Code and Codex. The biggest workflow win: give the agent ways to verify results instead of trusting its words.
Key practices
- Visual verification: dedicated pages for spells, cards, and maps let Claude capture frames via Playwright, inspect them, and iterate before human review.
- Gameplay bug detection: feed Claude a recorded player session and compare actual actions, inventory, and state changes against the narration — e.g. a merchant accusing the player of drawing steel when the player only owns a makeshift stick.
- Division of labor: the game maintains world state; the AI interprets actions and narrates. Keeping the two aligned remains most of the work.
Free to try at llmsaga.com.
Related event: Indie Dev Shares Three Years of Lessons Building an AI RPG with Claude Code(2 posts)→
More from coding & agent
- Dev floats idea of using agents to optimize Unreal Engine's source code — gajesh · 2026-11-03
- CleanWeb v2.8.4: SPA state parsing 95% faster, official LangChain/LlamaIndex packages — EstablishmentTough18 · 2026-11-03
- LangChain launches Managed Deep Agents: define agents as a directory of files — hwchase17 · 2026-11-03
- Designer builds AI website UX auditor on $1,500 budget, $7 spent so far — noya1979 · 2026-11-03
- Geiru Context Suite adds degradation protection for long MiniMax H3 video chains — roychodraws · 2026-11-03
- AI coding is 10–20x faster for the first 1–2 months, then complexity drags it back down — LadyAshBorg · 2026-11-03