EPITAPH

The description states that EPITAPH is an AI-driven historical narrative roguelite. It is built as a Steam game, with three distinct play modes:

• Chronicle: 8–15 minutes, offline auto-biography.
• Fate Cards: Reigns-style, controller-first, offline.
• Immersive: 90–180 minutes, full RPG driven by a live model.

Players are reborn as historical figures (e.g., Beethoven, Wu Zetian) at pivotal moments in their lives. They begin with:

• Their modern memory.
• The figure’s already-written future, represented as a deck of decaying prophecy cards.

Gameplay mechanics include:

• Two axes: HP and SAN (心神).
• Death is a real ending; SAN can deplete, causing the player to lose themselves.
• Epitaph Diff: a side-by-side comparison of the real encyclopedia entry and the life played, with changes highlighted.
• Points are earned and used to buy soul talents for reincarnation.

The game uses:

• A deterministic engine backed by SQLite and SHA-256 hash chains.
• AI only for narration; mechanical outcomes are committed before AI involvement.
• Tools like MCP server to allow AI agents to play the game as an agent, while maintaining deterministic control.

Not evidenced: actual gameplay footage, user reviews, monetization strategy, or any commercial data.

The author states that EPITAPH was built in response to what they call “AI games” that are merely chatbots in a costume — infinite text, zero stakes. The positioning is:

• A historical narrative roguelite.
• Grounded in foreknowledge as a resource, with history as pressure and death as real consequence.
• AI is used for narration only, not mechanical control.

The claim evolution shows:

• Initial inspiration: dissatisfaction with current AI games.
• Core positioning: a game where the player’s knowledge of history becomes a gameplay mechanic.
• Evolution toward a hybrid deterministic-AI architecture to ensure verifiability and replayability.

Not evidenced: any market positioning beyond the author's own description, or evidence of how this positions against competitors.

The description states that EPITAPH is aimed at:

• Players who enjoy historical narratives.
• Fans of roguelite mechanics.
• Gamers interested in AI-generated storytelling with a deterministic core.
• Players who value replayability and verification.

The game is designed for:

• Players who want to reincarnate as historical figures.
• Those who enjoy deep, meaningful choices with consequences.
• Users who are comfortable with offline play modes (Chronicle and Fate Cards).

Not evidenced: actual user data, demographics, or customer segmentation.

The description states:

• The game is built for Steam launch.
• It includes a ranked tier ("Official") and a UGC world pack system.
• There are three play modes, each with different fidelity levels.

No pricing information, monetization strategy, or revenue model is provided. The author mentions:

• A live hosted "Official" ranked tier.
• Workshop UGC world packs with fail-closed moderation.

Not evidenced: pricing tiers, in-game purchases, subscriptions, or any commercial structure beyond the vague mention of a Steam launch.

The project is built as a TypeScript pnpm monorepo using:

• React 19 + Vite
• Electron shell
• Fastify API
• SQLite with SHA-256 hash chains
• Ed25519-signed receipts for verification

AI components include:

• OpenAI Chat Completions and Responses API
• Anthropic, Gemini
• Codex (as in-game agent)
• MCP server to expose game tools as AI agents
• A2A character agents with A2UI surface

Key technical signals:

• Commit-then-narrate architecture.
• AI only narrates outcomes already committed by deterministic engine.
• Replayability and verification are core features.
• Security measures include:
  • Server-only providers
  • Secrets brokered through macOS Keychain
  • Electron fuses, ASAR integrity

Not evidenced: actual performance data, scalability, or deployment details beyond the architecture.

The description states:

• The project was built for the OpenAI 2026 hackathon.
• It is a single-person project (Leewaytor Li).
• The author claims to have accomplished:
  • A deterministic-AI architecture.
  • Full English/简体中文 switching mid-life.
  • One plugin package that works in both Codex and Claude Code.

No evidence of:

• User base or engagement metrics.
• Revenue or monetization.
• Public playtesting or feedback.
• Any form of traction beyond the hackathon submission.

The description does not provide any information about:

• Direct competitors.
• Market positioning relative to other roguelites or historical narrative games.
• AI game trends or how EPITAPH fits into them.

Not evidenced: competitive analysis, market size, or positioning against similar products.

• Single-person development: The project is built by one developer (Leewaytor Li), which raises questions about scalability and long-term maintenance.
• Niche genre: Historical narrative roguelites are a small niche; it’s unclear how the game will attract a broad audience.
• No commercial traction or monetization strategy: No evidence of revenue, users, or monetization beyond a Steam launch.
• High technical complexity: The architecture is complex (deterministic engine + AI narration), which may slow development or limit future expansion.
• Unproven market demand: There’s no evidence that the target audience exists in sufficient numbers to support a commercial game.

1. What is your plan for monetization beyond a Steam launch?
2. How do you intend to attract and retain players in a niche genre like historical roguelites?
3. Do you have any plans for community or UGC features beyond the Workshop packs mentioned?
4. How do you plan to scale beyond a single developer?
5. What is your roadmap for expanding the number of historical figures or play modes?
6. Have you considered how AI-generated content will be managed in terms of quality and consistency?
7. Are there any partnerships or distribution strategies planned with Steam or other platforms?