Project story

North Star began with a simple problem: AI coding sessions could do ambitious work, but the direction, proof, and next step were easy to lose between conversations.

Early versions explored project memory, mission tracking, schedules, and an operator dashboard. Real use exposed the important gap. More panels did not automatically make the system easier to understand. Internal terms, duplicated test records, and disconnected routes could make a truthful system feel opaque.

The current direction treats North Star as one connected command system:

• Briefing explains what matters now.
• Work shows readable tasks instead of queue machinery.
• Projects preserve long-running direction.
• Schedules make future work visible.
• Ask North Star provides conversation without hiding the task model.
• History keeps the proof and recovery path.

The public preview tells that story through the observatory and the Plan, Run, Review, Remember route. These docs turn the same story into a practical field manual.

The principle underneath both surfaces is simple: the point is not to automate everything. The point is to know what changed, why it changed, and where useful work should go next.

Read the North Star case study or open the public preview.