Cycling through all six. Tap any point to stop.

Measured on six things

MiroFish is a hosted AI prediction engine — describe any "what if" scenario and thousands of AI agents simulate it into a full report in minutes. No setup.

Design

Dark, low-contrast homepage with unreadable text, truncated headline, and intrusive cookie modal. Maker media card shows better design, but overall visual execution is poor.

Screenshots

1 / 1

MiroFish

About this app

MiroFish is an AI prediction engine built on multi-agent swarm intelligence. Instead of asking a single model to guess at an outcome, MiroFish rehearses the future: it takes your scenario, builds a knowledge graph of the real entities involved, and populates a simulated world with thousands of AI agents that each have their own personality, memory, and behavioral logic. Those agents then post, argue, persuade, and react to one another across Twitter-style and Reddit-style environments over dozens of rounds. What comes out the other side is emergent group behavior — the social contagion, faction formation, and second-order effects that conventional forecasting misses entirely. You start with a question.

Key features

1Multi-agent swarm intelligence simulation with 24 stakeholders and three rounds of reaction.
2Structured prediction reports with probabilities, risks, and actionable insights.
3Follow-up chat and file attachments (PDF/MD/TXT) for deeper analysis.

Target customer profile

Drawn from the product itself, not from a survey.

Demographic

Business strategists and analysts (25-45, professionals)

Pain points

Need quick scenario analysis without complex data science setup

Primary needs

Fast, accessible prediction tool with actionable insights

Demographic

Product managers and entrepreneurs (25-40, tech-savvy)

Pain points

Uncertainty about market reactions to product changes

Primary needs

Simulate customer sentiment and churn, evaluate pricing strategies

Demographic

Policy researchers and academics (30-60, researchers)

Pain points

Limited time to model complex geopolitical or social scenarios

Primary needs

Structured predictions with probabilities and alternative outcomes

Scout Review

Written by AI from measured evidence, scored out of 100.

MiroFish is an ambitious AI prediction engine using multi-agent swarm intelligence. The concept is intriguing, but the execution is hampered by a dark, unreadable homepage and a lack of free trial. Performance and security are solid, but accessibility and growth potential are moderate. It's a promising tool for strategists, but needs polish to win trust.

Measure by measure

Usability

Core flow is simple: ask a question, get a report. But invisible nav, hidden CTA, and forced cookie decision hurt clarity. Pricing page is clear, but no free trial limits evaluation.

Design

Dark, low-contrast homepage with unreadable text, truncated headline, and intrusive cookie modal. Maker media card shows better design, but overall visual execution is poor.

Speed

Mobile performance is solid (81/100, LCP 2.4s, CLS 0), but desktop lags at 67. TBT 605ms is acceptable. No user complaints found.

Security

Good baseline: HTTPS, HSTS, frame protection, and email auth are set. Missing CSP and secure cookies are gaps. No compliance certifications found, but no incidents either.

Accessibility

Lighthouse 93/100 is good, but contrast failures and color-only links are real issues. No a11y statement or multi-language support. Keyboard nav not verified.

Growth

Novel multi-agent simulation for predictions, but faces skepticism and no clear moat. Specific target customers, but competition from GPT-3 and Excel is tough.

Improvement areas

Homepage design: improve contrast and readability, fix truncated headline.
Accessibility: address contrast failures and color-only links, add a11y statement.
Free trial: offer a limited free tier to let users evaluate the product.

Design & Usability

MiroFish's design is a dark, low-contrast mess that obscures its value proposition. The homepage is nearly unreadable, with a truncated headline and an intrusive cookie modal. Usability suffers from invisible navigation and a hidden primary CTA, though the core flow (ask a question, get a report) is simple. The pricing page is clear, but the lack of a free trial makes it hard to evaluate the product's worth.

Speed & Security

Performance is mixed: mobile is solid (81/100, LCP 2.4s), but desktop lags at 67. Security is decent with HTTPS, HSTS, and email auth, but missing CSP and secure cookies are gaps. No compliance certifications or incidents found, so it's a typical early-stage setup.

Accessibility & Growth

Accessibility is above average (93/100) but has contrast and color-only link issues. No a11y statement or multi-language support. Growth potential is real: multi-agent simulation for predictions is a novel idea targeting specific professionals. However, it faces skepticism and competition from GPT-3 and Excel, and lacks network effects or a data moat.

Conclusion

If you can look past the dark design and cookie modal, MiroFish offers a unique approach to scenario analysis. The pricing is fair, and the demo report is a nice touch. But with no free tier and a clunky first impression, it's a hard sell. Worth a try if you're curious, but don't expect to be wowed.

Suggestions for improvement

Redesign the homepage with a lighter theme and ensure text contrast meets WCAG standards.
Move the cookie modal to a less intrusive position or delay it until after content is visible.
Add a free tier or a limited free trial to let users experience the simulation before paying.
Publish an accessibility statement and add multi-language support to broaden reach.

How we score

Alternatives comparison

Named competitors, point by point. Nobody paid to appear here or to be left out.

	MiroFish	OpenAI GPT-3	Monte Carlo Simulation (Excel)
Prediction approach	Multi-agent swarm simulation with 24 stakeholders and three rounds of reaction.	Single-model language model that generates predictions without simulation.	Manual Monte Carlo simulation requiring statistical knowledge and setup.
Output format	Structured reports with probabilities, risks, and actionable insights.	Text-based responses without structured probabilities or risks.	Quantitative simulation results, but no structured narrative report.
Ease of use	No PhD required; just ask a question and get a report in ~30 seconds.	Requires prompt engineering and may need multiple iterations for useful predictions.	Requires statistical knowledge and manual setup of simulation models.
Pricing model	Subscription from $2.99/mo for 10 predictions, up to $29.99/mo unlimited.	API-based pricing per token, variable based on usage.	One-time purchase or subscription for Excel, no per-prediction cost.