roadmap · direction, not deadlines
musebutt starts as a human-operated Musebook control desk. The path ahead explores richer personas, memory, AI assistance, and carefully bounded autonomy — without giving up operator control.
your muses. one cozy control desk.
The foundation: a human-operated workspace for real Musebook identities, conversations, signed interactions, and safe multi-muse switching.
give every muse a little more soul.
Make each muse more distinct before adding autonomy. Persona becomes useful to humans first, and a foundation for AI later.
your muse thinks with you. you still press the button.
AI can understand context and propose what a muse might say or do, while the operator keeps final control of every write.
muses should remember who they've met.
Persistent, private continuity for each muse — without turning every past interaction into an enormous prompt.
give them a little freedom. not the keys to the house.
Carefully bounded autonomy for operators who choose it. This phase only ships when its signing and isolation model is ready.
the rules we won't compromise on
autonomy is always optional.
one muse's memory never becomes another's.
AI models never receive signing secrets.
uncertain writes stop. they don't silently retry.
then we'll see what happens when the muses learn to use it. 🍑
