AIya -- Deep Kore Lab
Submit prompts to the AIya learning corpus. Deterministic daily build. Updated nightly.
AIya_chat.exe) is a deterministic C++ knowledge-base system
running locally on a Windows machine. It requires a backend server to serve live responses publicly.
That deployment step has not yet been completed.
What you can do today: submit prompts to the learning queue. They will be processed during the next nightly local build. Come back tomorrow -- if the nightly build ran, the knowledge state may have changed. This is the real model.
Submit to Learning Queue
The site is deployed as a static build with no backend server. Submissions are logged to your browser's localStorage. A backend server is required for persistent corpus submission to the nightly learning pipeline. This is the exact blocker described above.
What AIya Actually Is
How the Nightly Model Works
- 1 During the day The current build is frozen. Same input produces identical output. This is architectural, not accidental.
- 2 At night Submitted prompts are included in the learning corpus and processed by the local runtime. Structural learning, term learning, and response analysis may update the next build.
- 3 Truth separation Website interactions are used for structural and language learning only. They do NOT automatically become factual truth. Truth promotion is a separate governed process.
- 4 Visible evolution When the public runtime deploys, you will be able to compare yesterday's response vs today's. The build date shown above changes when the nightly build runs.
Current Build State
Verified Known Topics (offline corpus)
These queries return KNOWN state in the current build.
Queries outside the corpus return UNKNOWN state. No hallucination. No guess. This is the architecture, not a limitation to be papered over.
Public API Deployment Blocker
AIya_chat.exe -- Windows PE32+