How it works
A closed library. Real answers. No open-web guessing.
The thing that makes us different is not the model. It is what the model is allowed to look at.
Step 1
A closed library
We build and maintain a curated corpus of real operator experience: interviews, transcripts, vetted writing. The model only looks here.
Step 2
Retrieval, not guessing
When you ask, we retrieve the most relevant passages from the library and pass them to the model as the only source of truth.
Step 3
An answer with context
The model writes a plain-English answer, biased toward what worked for shops like yours. You get the answer and the why.
Step 4
Refusal when we do not know
If the library does not cover your question, we say so. We do not fall back to the open web to invent something.
Common questions
Which AI model do you use?
We use frontier models from established providers. The model matters less than the corpus. A smaller model with the right corpus beats a bigger model with the wrong one for this use case.
Why not use the open web?
Because the open web is where hallucinations come from. Marketing pages, content farms, and outdated forum posts are not how you want decisions made.
Do you train on my conversations?
No. Your conversations are yours. See our trust pages for the full data handling policy.
Have a question of your own?
Ask it in the chat. We answer from real operator experience, or we tell you the library does not cover it yet.