What It Does
FutureSearch is an AI forecasting and research platform that helps users estimate future outcomes with probability, numeric, date, categorical, conditional, and decision-based forecasts. It combines forecasting with structured research workflows, making it suitable for analysts, investors, researchers, and teams that need evidence-backed predictions rather than general AI conversations.
Unlike a standard chatbot, FutureSearch is built around forecasting as the primary workflow. Users can submit a question, choose an effort level that affects cost, review supporting research, and integrate the forecasting engine into AI tools like Claude, Cursor, and ChatGPT Codex through MCP.
A typical workflow starts with asking a forecasting question, generating supporting research, reviewing the forecast with evidence, and sharing or integrating the results into an existing AI workflow.
At a Glance
| Field | Details |
|---|---|
| Category | AI forecasting and research platform |
| Forecast Types | Probability, numeric, date, categorical, conditional, decision |
| Pricing Model | Pay per forecast ($0.15 to $2 per question) |
| Free Trial | $20 free credits |
| AI Integrations | Claude, Cursor, ChatGPT Codex |
| API | Available |
| Primary Strength | Research-backed forecasting workflows |
Key Features
- Generate multiple forecast types from a single research workflow.
- Combine forecasting with multi-agent research teams.
- Connect through MCP to Claude, Cursor, and Codex.
- Access forecasting through an API for custom workflows.
- Choose forecast effort levels based on cost and depth.
- Review published trading and forecasting performance.
Best For
- Estimating future business or market outcomes.
- Researching startup ideas before making decisions.
- Adding forecasting capabilities to AI coding workflows.
- Building evidence-backed probability estimates.
- Running repeatable forecasting through an API.
Pros & Cons
| Pros | Cons |
|---|---|
| Supports multiple forecasting formats | Forecasts still involve uncertainty |
| Pay-per-use pricing suits occasional users | Costs increase with higher-effort research |
| MCP integrations fit existing AI workflows | Not designed as a general productivity assistant |
| Public evaluation and trading results add transparency | Advanced workflows may require technical setup |
Alternatives & Comparisons
| Alternative | Best For | Key Difference |
|---|---|---|
| Metaculus | Community forecasting | Human crowd predictions instead of AI-first workflows |
| Perplexity AI | Research and answers | Strong research but not forecasting-focused |
| Polymarket | Prediction markets | Real-money market signals instead of AI-generated forecasts |
| Claude with MCP | AI productivity | Gains forecasting through FutureSearch integration |
FutureSearch occupies a specialized space between AI research assistants and prediction platforms. It is a stronger fit when forecasting itself is the deliverable rather than simply answering questions.
Frequently Asked Questions
Can FutureSearch predict different kinds of future events?
Yes. It supports probability, numeric, date, categorical, conditional, and decision forecasts.
Does FutureSearch work with Claude?
Yes. It integrates with Claude.ai and Claude Code through MCP connectors.
Is there an API?
Yes. FutureSearch provides an API for developers and custom workflows.
How is pricing structured?
Forecasts are priced individually, typically ranging from $0.15 to $2, depending on effort.
Can I try FutureSearch for free?
Yes. New users receive $20 in free credits.
Overall Rating
| Category | Rating |
|---|---|
| Performance | 4.8/5 |
| Ease of Use | 4.5/5 |
| Feature Set | 4.7/5 |
| Workflow Fit | 4.8/5 |
| Overall | 4.7/5 |
FutureSearch is a strong choice for users who need structured forecasting backed by research, especially when integrated into existing AI workflows.
Final Verdict
- Best for: Research-driven forecasting and decision support with AI integrations.
- Avoid if: You only need a general AI chatbot for everyday questions.
- Biggest strength: Multiple forecast types backed by structured research workflows.
- Best alternative: Metaculus for community-driven forecasting instead of AI-generated forecasts.
