Demo · run it on your AI

Ask this build history with your own AI

The public demo memory “AiAkiv-Roguelike” holds the full design, implementation, bugs and retrospectives of the turn-based roguelike Delve — and its repo and design docs are public too. Once your Main points at the demo project, drop the questions below into your AI: it searches the memory and the public git together.

Let your AI guide the exploration

Give the sentence below to the AI you just connected. It reads /llms-tour-demo.txt, first confirms you're on the demo memory, then walks the build history one question at a time. Stop anytime. (On ChatGPT — a new chat starts with the connector OFF. Click + in the composer and pick AiAkiv, or start from “Try in chat”.)

Use AiAkiv's mweft_guide(topic="demo") to fetch the demo walkthrough, then take me through this demo memory one question at a time. First check which memory I'm on, then explain as if to a first-timer.

Public repo & docs (the AI fetches these): https://github.com/rawdev/delve

Questions to throw

Tap a chip to replay an example answer and the causal path the memory stitched together. In practice your AI searches the demo memory (+ public git/docs) live. The first question confirms which memory you're on.

Tap a question

›

※ Example answers — the real ones are generated live by your AI.

How to run it on your AI

1

Point Main at the demo

like this →Console → Projects → switch to AiAkivRogueLike

Switch your save target (Main) to the demo project so your AI reads this public read-only memory. It's read-only, so you can't break the demo. New here? Set up the connection first.

2

Drop the questions in

like this →Copy a chip question into your AI

Ask the questions above verbatim. The AI runs mweft_search on the memory and, when useful, fetches the public git/docs too.

3

Dig into causality

like this →Why that decision? · What was rejected?

Follow the decisions and bugs in the answer and keep asking “why?”. Unlike keyword search, memory stitches cause to effect.

Not connected yet? Start with the usage tour.

To start on your own account, create a team and project in the console.

Open console →

Machine version for AIs: /llms-tour-demo.txt