FutureHouse, a nonprofit backed by former Google Chief Executive Officer Eric Schmidt that’s building artificial intelligence agents for scientific research, today announced the launch of FutureHouse Platform, a web-based and application programming interface-accessible suite of super-intelligent AI agents designed to accelerate scientific discovery.
The FutureHouse Platform has been designed to handle the overwhelming volume of scientific literature and data that scientists deal with when doing scientific research. The platform introduces four specialized agents, Crow, Falcon, Owl and Phoenix, to assist with tasks like literature synthesis, hypothesis generation and experiment planning.
The agents have been benchmarked to outperform human researchers in key tasks such as literature retrieval and synthesis. According to the company, the platform demonstrated higher precision and accuracy than PhD-level researchers in head-to-head evaluations of scientific search and summarization.
Of the agents, Crow serves as a general-purpose agent that delivers concise, scholarly answers and is optimized for API integration. The second agent, Falcon, offers in-depth literature reviews and has access to large-scale databases such as OpenTargets, which allows it to analyze more scientific content than any other comparable agent.
The third agent, Owl, answers whether a specific experiment or study has been conducted before, helping researchers avoid redundant work. The fourth agent, Phoenix, assists with experimental chemistry planning through specialized tools.
AI agents are hardly new in 2025, but where the agents get interesting is that, unlike general-purpose AI, they’ve all been built from the ground up for science. The agents can assess source reliability, track citations and surface methodological concerns by working with full-text papers rather than relying solely on abstracts or summaries.
The new platform also emphasizes transparency and reasoning, using a multi-step process that allows users to see exactly how conclusions are drawn. Doing so helps researchers validate findings and improves trust in AI-assisted workflows.
FutureHouse is making the platform accessible via both user interface and API, allowing labs and institutions to build custom research pipelines. Scientists can automate literature reviews, monitor new publications, or link the agents to experimental screening data for real-time context.
One of the first users of the platform has been the Michael J. Fox Foundation for Parkinson’s Research, which has been impressed with the results.
The platform “allows us to gather evidence from PD patient samples and preclinical in vitro and in vivo models in a highly structured and efficient manner,” said Gaia Skibinski PhD, director of discovery and translational research at The Michael J. Fox Foundation. “This work supports our Targets to Therapies initiative, launched last year, where a similar analysis was performed manually by a panel of scientific experts. It has been exciting to explore how advanced LLMs can complement and accelerate traditional scientific approaches.”
Image: FutureHouse
Your vote of support is important to us and it helps us keep the content FREE.
One click below supports our mission to provide free, deep, and relevant content.
Join our community on YouTube
Join the community that includes more than 15,000 #CubeAlumni experts, including Amazon.com CEO Andy Jassy, Dell Technologies founder and CEO Michael Dell, Intel CEO Pat Gelsinger, and many more luminaries and experts.
THANK YOU
