Altiverse

Altiverse is an AI simulation tool that forks a single decision into 2-4 alternate realities and runs them forward as living worlds for decision-makers and policy analysts. Instead of guessing at outcomes, you set a choice (a school policy, a workplace change, a healthcare protocol) and watch up to ~1,000 simulated people with their own personalities, relationships, and moods react differently across parallel timelines. The simulation is deterministic: it uses seeded randomness (Mulberry32 PRNG) so the same seed reproduces an identical world, which makes runs replayable and auditable. Default metrics include Focus, Grades, Trust, Stress, Burnout, Rule-breaking, Conflict, and Corner-cutting, and you can customize them to your scenario. Character cards show how a single person fares across every reality, causal-chain analysis links a policy to its consequence and second-order effects, and divergence visualization shows exactly when and why the worlds split. An optional AI layer generates scenarios, in-character thoughts, and written reports via an OpenAI-compatible client. Altiverse runs entirely in the browser (Vite + React + TypeScript), with no backend, no telemetry, and localStorage persistence, so nothing is uploaded; it even works offline without an API key. Results export to Markdown or PDF. It is open source under the MIT license, so it can be self-hosted and forked.

Policy decision simulation, second-order effect analysis, school policy testing, workplace change modeling, healthcare protocol evaluation, scenario planning, reproducible agent-based simulation