VivacityFIRST WEDGE / HYPOTHESIS TO VALIDATE
Start with teams building agents for simulated robotics.
The hypothesis: these teams currently stitch an agent to a simulator, write one-off state and replay tooling, then build task-specific checks around it. Vivacity aims to give each run one inspectable lineage: parent world → action → alternative branches → domain checks → committed state and replay.
This wedge and workflow description are a starting hypothesis, not a completed customer study or validated market claim.
CURRENT STATE / EVIDENCE
A working browser explainer; production integration is not claimed.
The public portfolio model runs deterministic two-body physics in the browser. It applies an action, compares branches, checks energy and surface clearance, commits a passing branch, replays its states, and exports a trace. The short agent-policy example is explicitly illustrative and does not call an LLM or the production API.
Next product proof point: one design partner completes its own simulator-backed task from initial state through action, branching, domain checks, commit, and reproducible replay. Measure setup time, replay success, and where the workflow fails. This is a target, not a completed milestone.
TRACTION / KEEPING THE CATEGORIES SEPARATE
A $10,000 grant, not investment or customer revenue.
Emergent Ventures awarded a $10,000 grant to build Vivacity; the program is administered by the Mercatus Center. It is grant support, not equity investment, a customer contract, or usage traction.
No dated user, pilot, or revenue figures are claimed on this brief. Design-partner conversations are the next evidence to turn into measured product usage.
Emergent Ventures
Mercatus CenterNEXT CONVERSATION
Bring us a real environment and a decision your agent needs to make.
We’re looking for design partners who can help test the first simulator-backed workflow and define the checks that matter.
Contact the Vivacity team