MiroFish

MiroFish is an AI simulation chat for scenario prediction. Instead of collapsing a complicated question into one confident answer, it builds a simulated world of actors, incentives and narratives, then hands you a structured prediction report you can keep questioning.

Ask a question, get a simulated scenario

You start in plain language: what happens to customer trust if we raise prices next quarter, which groups resist a new policy first, or how a launch narrative spreads in a skeptical category. Attach a PDF, Markdown or TXT file whenever you want the simulation grounded in a real memo, brief or research summary.

From seed material to a world you can question

Every run follows one continuous workflow: seed material, knowledge graph, agent simulation, prediction report and follow-up chat. The graph step extracts actors, relationships and factual anchors so agents reason from structure. Personas then interact across short-form and threaded social surfaces over multiple rounds.

A report that makes the next question obvious

Each report condenses emergent behavior into an executive summary, turning points, risk signals, narrative paths, confidence signals and follow-up questions, a rehearsal you can inspect before real customers, markets or institutions react. Treat it as exploratory decision support, not a guaranteed forecast.

Where teams use it

MiroFish fits decisions where human reaction loops matter: campaign testing, pricing reactions, policy stress tests, market narratives, demand and conversion planning.

Start in the browser

MiroFish runs in the web app with no install. Open the workspace at mirofish.homes, review current plans and run your first simulation.