Inspiration

Twitch communities build their own culture through inside jokes, memes, legendary moments, and recurring stories. But that history quickly disappears into the endless scroll of chat, leaving newcomers confused about what happened before they arrived.

We built LORE to give Twitch communities a living memory.

What it does

Viewers can nominate memorable moments directly from Twitch chat using !lore. LORE captures the message through Twitch EventSub and creates a proposed lore entry.

The streamer or moderator can review and approve the proposal. Once approved, it becomes part of the channel's permanent Lorebook.

Newcomers can then ask questions such as !ask What's the Banana Incident? and LORE uses the community's stored memories and evidence to explain the story.

LORE also provides a searchable Lorebook and an OBS-friendly overlay so community history can become part of the live stream itself.

The core loop is simple:

Viewer creates a moment → LORE captures it → Streamer approves it → Community remembers it → Newcomer asks about it.

How we built it

LORE uses Twitch EventSub for real-time chat events, a persistent worker for processing messages, Supabase for community memory, and OpenAI-powered responses for grounded questions and answers.

We used OpenAI Codex throughout development to build and iterate on the Twitch integration, worker, database layer, moderation system, AI interactions, APIs, Lorebook interface, OBS overlay, testing, and production deployment.

The project is deployed and connected to a real Twitch channel, demonstrating the complete flow from Twitch chat to persistent lore.

What's next

We want to make LORE available to any streamer through Twitch OAuth, with channel-specific Lorebooks, moderator accounts, automatic lore detection, and deeper integrations with streaming tools.

LORE turns fleeting Twitch moments into permanent community history.

Built With

Share this project:

Updates

Submission history