Personal memory and action
Liva listens, sees and interprets context. Flow keeps originals, Dock manages tasks and delegation, and Muse distils your day into new work. Nüwa tests how your corrections should change future responses.
Personal intelligence, agent negotiation and world execution.
Liva listens, sees and interprets context. Flow keeps originals, Dock manages tasks and delegation, and Muse distils your day into new work. Nüwa tests how your corrections should change future responses.
Turn an impulse to go out into a route, a shared interest and an outing both people choose.


Discover games, explore with friends or turn a local story into a playable world.
▶Travel with a little sun companion through dangerous places.
▶Move, fight and grow amid waves of enemies.
▶Swap elemental gems to trigger chains.
▶Find your rhythm between mountains and clouds.
▶Rescue a young creature or change the storm in a 90-second crisis.
▶Merge matching spirits and make room for the next drop.
▶Turn an idea into a business, meet people who remember, and live with twelve weeks of decisions.
▶Weave through swarms, collect upgrades and survive the night.
▶Find an angle and break through with a chain of rebounds.
▶Absorb spirit energy, evade rivals and grow into a dragon.
▶Time each drop and build upward through the sway.
▶Draw your territory and protect the route home.
▶Find the gap in the rotation and land each blade.
▶Read moving hazards and find your next crossing.
▶Collect more or leave in time? Each step asks a new question.
▶Rotate a cube and find connected groups of matching colors.
▶Defend the homefire and grow a home through combat and upgrades.

Choose an exploration template, write the goal and place clues. Save an AI revision as a candidate, playtest the rules and publish your version.
Start from a template →Flow contributes selected material and Seek contributes permitted local clues. Land handles design, playtesting and publishing. Automatic cross-product handoff is in development.
Flow is in iOS beta. Seek and Land are illustrated with current prototypes. The full cross-product journey is a design goal.

“Use the Long Street archive I found to open a route from the old station to the courtyard. I will supply the courtyard archive later. Negotiate the route only. Do not move or execute for us.”
Share the Long Street archive → I owe a courtyard archive
The approved first plan opened the route. The character stayed put and I retained one archive obligation.
“Open the same route. My partner will provide the courtyard field observation. I will share the existing Long Street archive. Negotiate only. Do not move or supply the material yet.”
Share the Long Street archive → My partner owes a field observation
After the first plan ran, submitting this stale alternative returned 409. Ownership stayed unchanged and the partner was not moved.
Products invoke the engines. After Fuxi commits an event, consented experience feeds evaluation and released capabilities inform future actions.
Memories retain their sources. CL and the three RSI levels produce candidates for independent, equal-budget testing before release or rollback.
Multimodal perception → temporal memory and relationships → goal and action proposal → runtime checks
A2A negotiates shared terms. After both parties consent, Fuxi commits. Each agent keeps its own private preferences.
State and legal action → reachability, risk and rollouts → compare with receipts → revise planning
Licensed assets → scene and rule IR → agent playtest and local edits → dependencies and attribution → load
A player in the field looks for clues. A remote partner works from different information. Both decisions affect the same session. Afterwards each person chooses what to keep. Open the web prototype or inspect the iPhone prototype screens.
Open prototype ↗Gameplay research (中文) ↗Proposed observation windows. Human retention, repeat visits and payment have no published measurements yet.