FutureHouse Platform

The FutureHouse Platform is a publicly available suite of superintelligent AI agents designed to accelerate scientific discovery for researchers everywhere. Accessible via both a web interface and API, the platform addresses the critical information bottleneck facing modern science — including the 38 million papers on PubMed, 500,000+ clinical trials, and thousands of specialized tools — by deploying AI agents with benchmarked superhuman literature search and synthesis capabilities.

Built from the ground up for science, FutureHouse's agents have been rigorously benchmarked and outperform all major frontier search models on retrieval precision and accuracy. In head-to-head literature search tasks, their retrieval and synthesis abilities have been experimentally validated as having better precision than PhD-level researchers. The platform is developed by FutureHouse, a registered 501(c)(3) nonprofit organization.

Available AI Agents

• Crow — A general-purpose agent that searches the scientific literature and provides concise, scholarly answers to questions. Crow is particularly well-suited for use via API.
• Falcon — Specialized for deep literature reviews, Falcon can search and synthesize more scientific literature than any comparable agent. It also has access to specialized scientific databases such as OpenTargets.
• Owl (formerly HasAnyone) — Specialized to answer the question "Has anyone done X before?", helping researchers quickly identify prior work in a given area.
• Phoenix (experimental) — A deployment of ChemCrow, an agent with access to specialized chemistry tools that assists researchers in planning chemistry experiments. Phoenix is released in the spirit of rapid iteration and may make more mistakes than the other agents.

Key Differentiating Capabilities

• Access to high-quality open-access literature: FutureHouse agents draw from a vast corpus of open-access papers and specialized scientific tools, enabling retrieval from specialist databases and automation of chemistry workflows.
• Source quality evaluation: Agents use a variety of methods to evaluate source quality, mirroring the approach a skilled researcher would take.
• Transparent, multi-stage reasoning: Agents employ a multi-stage process to consider each source in depth, and every user can inspect this reasoning to understand exactly how a conclusion was reached.
• Full-text paper access: Unlike tools limited to abstracts, FutureHouse agents can answer questions about experimental methods and study limitations found only in the full text of papers.
• API and web interface: The platform is built for scale, offering both a web interface and an API to facilitate and integrate into researcher workflows without requiring scientists to maintain their own agent deployments.

Example Use Cases

• Identify unexplored mechanisms in disease pathways: Use Falcon for background knowledge, Crow to identify key genetic associations, and Owl to pinpoint research gaps — all in minutes rather than weeks of manual literature review.
• Systematically identify contradictions in the literature: Task Falcon with analyzing conflicting evidence across hundreds of papers on controversial topics and highlight where additional experiments could resolve conflicts.
• Conduct critical analyses of experimental methods: Leverage full-text access to ask agents about experimental methods or study limitations not apparent from abstracts alone.
• Customize research pipelines via API: Research groups can build automated systems that continuously monitor new publications or conduct literature searches at scale to contextualize screening experiment results.
• Find known hits that bind to a target protein: Task Phoenix to propose hits to a protein target using existing data sources, with complex constraints on solubility, functional groups, or novelty.
• Reason about chemical space: Phoenix can assess compound novelty, estimate costs, predict reaction outcomes, and determine whether it is cheaper to buy or synthesize a given compound.

The FutureHouse Platform is available for free at platform.futurehouse.org. By chaining agents together at scale, scientists can greatly accelerate the pace of discovery. The platform is especially effective for challenges requiring detailed full-text literature analysis or the use of specialized chemistry tools via Phoenix.