Tempest

Tempest is a decision-intelligence workspace for founders and early-stage teams. Most AI tools are good at answering questions, but they still tend to lose context, forget why decisions were made, mix old and new information together, and give generic advice. Tempest is designed around a different idea: your company context should compound over time. It keeps track of what you are building, the decisions you make, why you made them, the assumptions behind them, and what happens afterwards. If you contradict an earlier decision, Tempest can notice it and ask what changed instead of silently overwriting the past. It can also challenge weak assumptions, help you decide what to do next, turn that into actionable steps, and continue from where you left off when you return with the results. Key points: Persistent startup and company context Decision history with reasoning behind each choice Contradiction detection Evidence-based challenges instead of constant agreement Clear next-step guidance Continuous workflow: decide → execute → return → learn Context that becomes more useful as the company evolves Planned integrations with tools such as GitHub, Stripe, Reddit and others, so Tempest can learn from real activity and outcomes instead of relying only on manual updates We’re currently preparing a small closed alpha with a limited group of founders. The main thing we want to test is simple: Does Tempest actually become more useful the longer a founder uses it? We’re looking for founders who already use AI in their startup workflow and are willing to test Tempest with real company problems, decisions, and execution.