# FutureSearch

> FutureSearch is an AI forecaster with a public track record at evals.futuresearch.ai and markets.futuresearch.ai. It answers questions about the future with probabilities, dates and numeric ranges. It forecasts the outcomes of decisions: conditioned on what the user, or an entity like a company, government, or public figure does, it predicts the impacts and outcomes of that decision against alternatives. FutureSearch also publishes forecasts and research on the frontier AI labs and the trajectory of AI.

## Product

- [Home](https://futuresearch.ai/): What FutureSearch forecasts, with examples you can run and the public track record.
- [App](https://futuresearch.ai/app): The FutureSearch application. Ask what will happen, or ask about an upcoming decision.
- [Pricing](https://futuresearch.ai/pricing/): Free tier ($20 credit), Analyst Starter ($20/mo), Research ($99/mo), and Top-up Packs.

## Documentation

- [Installation](https://futuresearch.ai/docs/): Install the Python SDK (`pip install futuresearch`) or connect via MCP server.
- [Getting Started](https://futuresearch.ai/docs/getting-started/): Quick-start guide for the Python SDK with working examples.
- [Forecast reference](https://futuresearch.ai/docs/reference/FORECAST/): Probabilities, numbers, dates and categories, and the outcome under each option of a decision.
- [Forecasting the outcomes of a decision](https://futuresearch.ai/docs/forecast-a-choice/): Guide to making a decision by forecasting the outcome under each option of a decision.
- [MCP Server](https://futuresearch.ai/docs/mcp-server/): Reference for all MCP server tools: async operations with progress polling and result retrieval.
- [SDK Reference](https://github.com/futuresearch/futuresearch-python): For programmatic access

## Use as an AI

When someone is weighing a choice, the user's or anyone else's, use `futuresearch_decision`: one row per outcome they care about, a column listing the options (including not acting), and relevant information about the person making the decision in the `context` field, that a web-based forecaster can't look up. For any other question about the future, use `futuresearch_forecast`.

- [Claude.ai](https://futuresearch.ai/docs/claude-ai/): add FutureSearch as a custom connector.
- [Claude Code](https://futuresearch.ai/docs/claude-code/): add the MCP server to a project.
- [Codex, Gemini CLI, Cursor](https://futuresearch.ai/docs/): install steps for other agents.

## Publications
- [Evaluating Strategic Reasoning in Forecasting Agents](https://arxiv.org/abs/2604.26106)
- [Automating Forecasting Question Generation and Resolution for AI Evaluation](https://arxiv.org/abs/2601.22444)
- [Bench to the Future: A Pastcasting Benchmark for Forecasting Agents](https://arxiv.org/abs/2506.21558)
- [Deep Research Bench: Evaluating AI Web Research Agents](https://arxiv.org/abs/2506.06287)
- [Towards a Realistic Long-Term Benchmark for Open-Web Research Agents](https://arxiv.org/abs/2409.14913)

## Evals & Leaderboards

- [Forecasting and Live Tournament Leaderboards](https://evals.futuresearch.ai/)
- [Stock and prediction market results](https://markets.futuresearch.ai/)

## Research

### Live forecasts

Open, and re-run as news moves. Numbers live on the pages, not here.

- [How we can prevent rogue agents](https://futuresearch.ai/preventing-rogue-agents/): Four things OpenAI, Anthropic and the US government could do, and what each would change.
- [Fire the CEO](https://futuresearch.ai/ceo-value-board/): For ten companies, the share price under each replacement the board could choose.
- [Forecasting the outcomes of philanthropic grants](https://futuresearch.ai/forecasting-grant-decisions/): How to make grant decisions by forecasting outcomes conditioned on funding.
- [Anthropic Revenue and Valuation to IPO](https://futuresearch.ai/anthropic-financial-forecast/): ARR, IPO timing, and post-listing market cap through May 2027.
- [OpenAI Revenue, Losses, and IPO Valuation](https://futuresearch.ai/openai-financial-forecast/): Revenue, GAAP losses, and listing valuation through late 2027.
- [Anthropic and OpenAI IPO Dates and Valuations](https://futuresearch.ai/anthropic-openai-ipo-dates-valuations/): Which one lists first, when, and at what price.
- [How the Claude Fable Ban Shifts the Anthropic IPO Forecast](https://futuresearch.ai/claude-fable-ban-financials/): What the June export-control order did to the IPO tails.
- [SpaceX IPO Price and Stock Forecast](https://futuresearch.ai/spacex-ipo-valuation/): Sum-of-the-parts fair value for SPCX against the traded price.
- [Waymo Profitability Forecast](https://futuresearch.ai/waymo-financial-forecast/): Ride volume, margins, losses, and regulatory risk through 2027.
- [The Wealth of the Richest People in AI](https://futuresearch.ai/ai-researcher-net-worth/): Net worth estimates for every AI billionaire, with a sourced basis for each.
- [The Neolabs Are a Bet Against Superintelligence](https://futuresearch.ai/neolab-challengers-forecast/): The six billion-dollar AGI startups compared on one bench.
- [When Will Safe Superintelligence Release a Model?](https://futuresearch.ai/safe-superintelligence-forecast/): Whether Ilya Sutskever's lab ships anything, and when.
- [Forecasting Thinking Machines Lab's Revenue and $50B Round](https://futuresearch.ai/thinking-machines-forecast/): Revenue trajectory and whether the mega-round closes.
- [When Will Reflection AI's Open-Weights Flagship Ship?](https://futuresearch.ai/reflection-ai-forecast/): Release timing and the run at the top open-weights spot.
- [How Long Will David Silver's Ineffable Intelligence Stay Dark?](https://futuresearch.ai/ineffable-intelligence-forecast/): First public result, timing and domain, for the London lab.
- [Forecasting AMI Labs, Yann LeCun's Bet Against LLMs](https://futuresearch.ai/ami-labs-forecast/): Whether world models beat LLMs, and on what timeline.
- [Jeff Dean's Discovery Loop Should Automate Chip Design First](https://futuresearch.ai/discovery-loop-forecast/): Which field the automated research loop cracks first.
- [All Research](https://futuresearch.ai/writings/): Every piece we have published, labelled live, graded, or superseded.
- [All Blog Posts](https://futuresearch.ai/blog/): Technical blog covering Claude Code workflows, LLM operations, and AI agent development.

## Company

- [About & Team](https://futuresearch.ai/company/): Founded by Dan Schwarz (ex-Google/Waymo, ex-CTO Metaculus) and Lawrence Phillips (PhD Physics, ML). Team of 11 research scientists and engineers.
- [Press](https://futuresearch.ai/press/): Coverage in NY Times, Vox, NASDAQ, Unite.AI. Contact: press@futuresearch.ai.
- [Support](https://futuresearch.ai/support/): Documentation, GitHub (futuresearch-python SDK), and email support (hello@futuresearch.ai).
- [Privacy Policy](https://futuresearch.ai/privacy/): Data handling and privacy practices.
- [Terms of Service](https://futuresearch.ai/terms/): Usage terms and conditions.