One scope per project. No mixed wires.
A README, the sources you trust, and every chat and draft scoped to it. The AI stops starting over and stops mixing up your projects.
Plays nicely with the tools you already pay for
repos
code + PRs
calls
Gong, Granola
docs
Notion, ADRs
tickets
Linear, Jira
Chat, mentions, research, and plans all inherit this scope
One workspace, everything inherits
Instructions, sources, and a running summary that the AI keeps current. Every chat and draft inside the workspace already knows the constraints.
A README the AI actually reads
Plain markdown. What the project is, what to look at, what to avoid, what's already decided. Update it as you learn, the AI stays current.
- Plain markdown, no rigid template
- A 'don't do this' section the AI honors
- A decisions log so the same call doesn't get re-argued
A canvas where every card cites its source
Drag observations off your Gong calls and docs onto the scratchpad. Group them into themes, frame an HMW, ask Kasava for solutions. Every card links back to the quote it came from.
- Observation → Theme → HMW → Solution, visualized
- Drop in calls, URLs, uploads, or notes directly
- Trace any proposed solution back to the evidence that produced it
Chat that doesn't make you re-explain
Every chat inside the workspace inherits the README and the sources. You don't have to type the constraints into every prompt. The AI already knows.
- Workspace context loaded automatically into every chat
- Answers cite the sources you added, not whatever the AI guessed
- README rules shape what the AI suggests
Make one workspace. Watch everything inherit.
Add a README and two sources. Every chat and draft inside already knows the constraints. Free to try. No card.
No credit card required