Stop getting
generic AI advice
on decisions that matter

Every Gavel citation links to a real, retrievable original source — a podcast episode, book chapter, article URL, or recorded talk. Click Show source on any quote to see the verbatim chunk and original artifact.

The fake-citation problem in generic LLMs (the 2023 NYT-covered lawyer case is the canonical one) isn't a hypothetical risk for us — retrieval-with-source is the foundation of how Gavel answers.

The retrieval layer is the product. ChatGPT generates from compressed training data and often hedges. Gavel retrieves specific named-expert positions from a curated corpus of 1,000+ frameworks before generating — and shows you the source.

The underlying model is interchangeable (we route across Claude, GPT-5.1, and Gemini 3). The corpus, the retrieval, and the "where experts disagree" synthesis are the moat. None of that lives inside ChatGPT.

Both of those are single-corpus tools — YC essays only, Lenny's content only. They're great if your question fits inside one person's worldview.

Gavel spans 8+ domains AND surfaces where named experts disagree across them — Patrick Campbell vs Jason Lemkin on freemium, Marty Cagan vs John Cutler on prioritization, April Dunford vs Andy Raskin on positioning. No single-corpus tool can do that.

Yes. Onboarding asks for your stage, business model, regulatory context, and team size. You can edit any of it any time — your context is visible, not black-box.

Every framework retrieval is filtered through your context. If a framework doesn't fit (e.g., a VC-stage growth playbook applied to a bootstrapped B2B), Gavel says so explicitly instead of forcing it.

Gavel surfaces named expert positions — not its own opinion. When experts disagree (which they often do), Gavel shows both sides, then tells you which one fits your specific context and why.

It doesn't hedge with "it depends." It picks a side — with citations — and explains the trade-off so you can override if you disagree.

Gavel tells you so — explicitly. "I don't have curated expert knowledge on this. Here's general guidance from related frameworks, clearly labeled as such."

Transparency over hallucinated confidence. This is a load-bearing principle — we'd rather lose a question than fabricate an answer.

Yes. The Skills page lists every expert, framework, and domain currently in the corpus — with sample frameworks viewable without an account.

If a specific expert you trust isn't there, tell us and we'll add them.

Your business context is private to your account. You can view and edit your entire memory at any time. We don't train models on your content.

Plus and Pro tiers include persistent memory. Free tier does not store context between sessions.