
FutureHouse Review, Pricing & Alternatives (2026)
Foundational AI Scientist research for biology and scientific discovery
pivoted: On November 5, 2025, management of the public platform moved to Edison Scientific. FutureHouse continues foundational AI Scientist research in biology, while Edison operates the spun-out platform through Edison Playground and restricted Kosmos access.
Last updated: October 26, 2026
What is FutureHouse?
FutureHouse is a nonprofit research organization developing AI Scientists for biology and scientific discovery. Its former hosted platform moved to Edison Scientific in November 2025, where Literature, Analysis, and Precedent agents remain available through Edison Playground. Researchers receive limited free access, while new Kosmos subscriptions are forthcoming and current public pricing is not confirmed.

FutureHouse is a nonprofit research organization focused on AI Scientists for biology and scientific discovery. The organization launched its public agent platform in 2025, giving researchers web and API access to specialized scientific tools. Its early platform included Crow, Falcon, Owl, and Phoenix, followed by Robin as a broader research workflow and Finch for data analysis. On November 5, 2025, management of the hosted platform moved to Edison Scientific, a commercial spinout. FutureHouse retained foundational AI Scientist research in biology, while Edison took responsibility for product access and commercial deployment. A May 19, 2026 update confirms that core research agents remain available through Edison Playground. FutureHouse is therefore best understood as the nonprofit research group behind the technology, not the current platform operator.
FutureHouse's former agents divided scientific work into specific tasks. Crow searched literature and returned concise scholarly answers. Falcon performed deeper reviews using open access papers and specialist databases, including OpenTargets. Owl, later identified in Edison materials with the Precedent function, checked whether an experiment had already been performed. Phoenix, an experimental ChemCrow deployment, helped plan chemistry experiments with specialized tools. Robin combined literature search, hypothesis generation, candidate ranking, and optional Finch data analysis. A typical workflow accepted a disease or research question, generated candidate assays or therapies, ranked findings, and saved reviews, summaries, CSV files, and analysis outputs. The former platform also supported custom literature pipelines through web and API access.
Edison describes Kosmos as a deep research system that maintains a research objective across long investigations and combines information from many agent trajectories. The company says individual runs can read 1,500 papers and execute 42,000 lines of analysis code. Kosmos is associated with the November 2025 platform transition, but Edison’s May 19, 2026 update says new Kosmos access is restricted while the company expands capacity. Edison Playground currently keeps the Literature agent, formerly Crow, the Analysis agent, formerly Finch, and the Precedent agent, formerly HasAnyone or Owl. The same update gives all users 10 free monthly credits for those agents. Robin remains an open source research reference with Python, Jupyter Notebook, and Docker workflows, but platform agents and model providers require separate access.
Edison’s current free access includes 10 monthly credits for core Literature, Analysis, and Precedent agents in Edison Playground. Academics who previously received 650 credits per month are moving to a one-time 2,000-credit grant, while Edison says it plans to announce an updated academic tier later. Founding Kosmos subscriptions are no longer available. Edison previously announced Kosmos at $200 per run, equal to 200 credits at $1 per credit, but that launch offer is not a current public subscription plan. The company says new Kosmos subscriptions are forthcoming, with access currently limited as capacity expands. The public update does not provide confirmed current subscription prices, quotas, or API rates. Academic grant programs have separately offered 20,000 credits for selected projects.
The main limitation is the separation between FutureHouse research and Edison’s hosted products. Kosmos can conduct large investigations, but Edison says long runs can follow statistically significant yet irrelevant findings, so researchers must verify results. Semantic Scholar is better for conventional scholarly search, Elicit emphasizes evidence tables and review workflows, and Ai2 Asta covers literature, datasets, and theory exploration. FutureHouse’s earlier agents remain useful references for scientific agent design, while Edison Playground is the practical access point for current tools. Researchers following nonprofit AI Scientist work should choose FutureHouse. Users seeking current hosted research agents should choose Edison Scientific, while teams needing focused literature review should choose Semantic Scholar or Elicit.
Who is FutureHouse best for?
- Biology researchers following foundational AI Scientist work.
- Academic users needing limited free scientific-agent access.
- Developers studying literature and discovery agent architectures.
- Teams evaluating Edison’s research workflows before wider Kosmos access.
FutureHouse Pricing (2026)
Pricing model: Freemium. All prices in USD.
| Plan | Price | What you get |
|---|---|---|
| Edison Playground | $0 | All users receive 10 free credits per month for Literature, Analysis, and Precedent agents. The May 2026 update does not publish additional free-tier quotas. |
| Academic grant | $0 | Academics previously receiving 650 monthly credits are moving to a one-time 2,000-credit grant. Edison says an updated academic tier is planned, but no current recurring quota is published. |
| Kosmos subscriptions | Not publicly listed | Founding subscriptions are no longer available. Edison says new subscriptions will be announced, while current Kosmos access is restricted as capacity expands. |
| Enterprise | Custom | Edison offers enterprise access and support by contact. Public pages do not list enterprise quotas or rate limits. |
Pros and Cons
Pros
- FutureHouse publishes research behind its scientific agents.
- Edison Playground retains Literature, Analysis, and Precedent agents.
- Robin’s code is available under the Apache 2.0 license.
- Academic users receive free Edison credits.
Cons
- Platform responsibility moved from FutureHouse to Edison Scientific.
- Kosmos access and subscription pricing are currently restricted or unpublished.
- Long investigations can follow irrelevant research paths.
Who should look elsewhere?
- Users expecting FutureHouse to operate the former hosted platform.
- Researchers needing deterministic answers from short chatbot prompts.
- Teams requiring confirmed public Kosmos subscription pricing.
Technical Details
Frequently Asked Questions
What is FutureHouse now?
FutureHouse is a nonprofit research organization developing AI Scientists for foundational biology research. It no longer operates the public platform launched in 2025. On November 5, 2025, platform management moved to Edison Scientific, which operates Edison Playground and controls access to the spun-out Kosmos product.
Is FutureHouse free to use?
FutureHouse’s former platform is no longer the current access point. Edison Playground gives all users 10 free monthly credits for Literature, Analysis, and Precedent agents. Academics who previously received 650 monthly credits are moving to a one-time 2,000-credit grant. Edison says a revised academic tier is planned.
How much does Kosmos cost now?
Edison previously launched Kosmos at $200 per run, or 200 credits at $1 per credit. That launch offer is not a current public subscription plan. Founding subscriptions are no longer available, and Edison says new Kosmos subscriptions will be announced. Current restricted access does not include a confirmed public price.
How does FutureHouse compare with Semantic Scholar?
FutureHouse focuses on AI Scientist research and specialized workflows for biology and scientific discovery. Semantic Scholar is a conventional scholarly search and discovery service with broader day-to-day literature coverage. Edison Playground is more suitable for agent-assisted analysis, while Semantic Scholar is better for quickly locating and filtering academic papers.
Does FutureHouse train on user data?
FutureHouse’s privacy policy says it may use personal information for data training, service development, research, and content improvement. The policy applies to FutureHouse services and websites. Edison operates the spun-out platform separately, so users submitting research data through Edison Playground or Kosmos should review Edison’s applicable privacy terms before uploading confidential material.
Can I use FutureHouse agents commercially?
Commercial hosted access belongs to Edison Scientific after the November 2025 transition. Edison provides enterprise contact options, but current Kosmos subscriptions and commercial quotas are not publicly listed. Robin’s code is Apache 2.0 licensed, although that license does not grant rights to proprietary models, datasets, platform services, or third-party outputs.
Best FutureHouse Alternatives in 2026
These tools do the same job as FutureHouse, with different pricing, limits and trade-offs.

AI-assisted search and discovery for scientific literature
AI workflows for literature search, review, and research reports

Evidence-linked search and synthesis for academic literature

Cited literature search and data analysis for scientific work