The future of agentic AI depends on trustworthy information infrastructure. AI agents can search, reason, and generate answers, but need the right evidence to do so well. That is why we built the Elicit API and MCP server. In this live demo, Elicit engineer Panda walks through how developers and researchers can: • Search millions of academic papers and clinical trials through the Search API • Generate evidence-backed reports with Reports API and Systematic Review API • Run systematic literature reviews directly from tools like Claude and ChatGPT • Build custom applications powered by scientific evidence Watch the recording to see how Elicit brings trusted evidence into the next generation of AI workflows: https://lnkd.in/gmgs89JU
Elicit
Research Services
Oakland, CA 8,891 followers
Elicit helps researchers be 10x more evidence-based
About us
Elicit, the AI research assistant, helps you automate time-consuming research tasks like summarizing papers, extracting data, and synthesizing your findings. We're a public benefit company with a mission to scale up good reasoning. We want machine learning to help as much with thinking and reflection as it does with tasks that have clear short-term outcomes.
- Website
-
https://elicit.com/
External link for Elicit
- Industry
- Research Services
- Company size
- 11-50 employees
- Headquarters
- Oakland, CA
- Type
- Privately Held
- Founded
- 2023
Locations
-
Primary
Get directions
Oakland, CA, US
Employees at Elicit
Updates
-
We are one day away from a live demo of the Elicit API and MCP server. We'll cover: - Search API for searching 138 million papers and 545,000 clinical trials - Reports API for generating cited reports - Systematic Review API for running a systematic review end-to-end - MCP server for using Elicit inside Claude and ChatGPT Register to attend live or get a copy of the recording: https://lnkd.in/g47XBm8x
-
-
AI agents are taking on more research work and need access to reliable evidence. That's the problem we're trying to solve with Elicit API and MCP server. We're demoing these capabilities live next Wednesday. We’ll cover: - Searching 138 million papers and 545,000 clinical trials through the Search API, which showed the highest paper recall against four popular search tools on the BioASQ benchmark. - Turning a research question into a fully cited report with the Reports API. - Running a systematic review end-to-end with the Systematic Review API, with control over search strategy, screening, extraction, and reporting. - Connecting Elicit to Claude and ChatGPT over MCP and showing what evidence-grounded agents look like in practice. Title: AI Agents Grounded in Evidence: A Live Demo of the Elicit API and MCP Server Presenter: Panda, Software Engineer at Elicit Date and Time: Wednesday, July 29, 2026, 10am PT Register: https://lnkd.in/g47XBm8x
-
-
The Elicit API and MCP are now generally available. To support this launch we rebuilt Elicit's search from the ground up and tested whether it actually was better. We picked BioASQ, a peer-reviewed biomedical benchmark to measure our performance. We compared Elicit against Consensus, Semantic Scholar, OpenAlex, and Google Scholar. Elicit had the highest recall at every result depth we tested from 10 to 200 results. At 50 results, Elicit retrieved 60.3% of the papers experts deemed sufficient to answer their question, versus 47.4% for the next-best system. Read the full write-up: https://lnkd.in/gcYtw52Y
-
-
Over the past year, Orion Pharma, a leader in human and veterinary medicines, has gone deep with Elicit. The team began using the platform in the early days of Reports and Systematic Literature Reviews, and more recently expanded that work with the launch of our Research Agent. Serhii Vakal, PhD, Lead ML/AI Scientist at Orion Pharma, shared a compelling story about how he used Elicit to search, collate, extract, and interpret large volumes of scientific literature far more efficiently than manual review. Work that previously required going paper by paper can now be completed in a fraction of the time, helping Orion Pharma evaluate more targets with the same team and reach each assessment earlier. We're excited to see how Orion Pharma continues to build on this work as Elicit supports its early discovery workflow. Read the full case study: https://lnkd.in/gxrTDmaw
-
-
Researchers do not need to become copyright experts to use AI responsibly. But they do need a practical way to understand what content they can use and how. In our latest fireside chat, Sheila Leunig, General Counsel at Elicit, and Zen Jelenje, Founder of Nascent Studio and former Data Solutions Lead at Elsevier, discussed how researchers can navigate copyright and data rights as AI becomes a larger part of scientific research. They covered: - Why access to an article or dataset does not automatically grant permission to use it with AI - How copyright rules, publisher licenses, the way content is accessed, and its intended use can lead to different answers - Why “open access” does not mean unrestricted use - Why trustworthy scientific AI is a shared responsibility across researchers, institutions, publishers, data providers, and AI platforms Watch the full fireside chat: https://lnkd.in/gJipRfjM
AI, Copyright, and Data Rights: A Framework for Scientific Research
https://www.youtube.com/
-
Aletheia AI has launched ReBind, a personal AI scientist designed to help clinicians investigate treatment options for individual patients. The Elicit API powers ReBind's evidence layer, helping it search scientific literature, retrieve supporting references, and show where the evidence supports or contradicts a proposed option. This is the role we want Elicit to play in AI-powered research systems: helping teams find and verify the evidence behind a conclusion, while keeping experts in control of the final decision. Learn more at https://lnkd.in/gN3Cf7Tk. Congratulations to the Aletheia AI team on the launch!
Every patient is unique. Their care should be too. Yet the standard of care is built for the general case: the same pathway, the same guideline, the same protocol, regardless of who's in front of it. But no patient is a general case. Two people with the same diagnosis can have entirely different biology, and the treatment that works for one can fail the other. Today, we're launching ReBind. ReBind is a personal AI scientist for every patient. It continuously monitors patient data, identifies meaningful changes, investigates possible interventions, and presents clinicians with the evidence, uncertainty, and reasoning behind each option. The clinician remains the decision maker throughout. It combines statistical monitoring, scientific literature, molecular modelling, and regulatory information in one traceable system, distinguishing real signals from noise, then evaluating potential options against the patient's specific biology. The evidence layer is powered by Elicit's API, which lets ReBind search published literature and retrieve traceable references for each treatment option, checking whether existing research supports, challenges, or contradicts a proposed approach. Boltz predicts how candidate medicines may interact with a patient's specific protein target, flagging options worth investigating rather than proving they'll work. Amass provides the regulatory layer, checking approval records and safety information before an option can progress. Together, these connect published evidence, molecular prediction, and regulatory verification in one system where every conclusion comes with its sources attached. If the literature contradicts a prediction, or a regulatory check fails, ReBind moves to the next option and explains why. The result isn't an autonomous prescription. It's a fully cited scientific investigation that helps a clinician make a more informed decision. Watch the video to see ReBind in action through a clinical case study: a simulated patient with real suggestions, evidence, citations, and regulatory checks generated live by our agentic backend. Over time, working with the right clinical and research partners and with the right safeguards, the knowledge generated will support research into drug discovery and new treatment strategies. Thanks to the Elicit team for their API and support around this launch. We are Aletheia AI. #AletheiaAI #ReBind #Elicit #Boltz #Amass #HealthAI #ClinicalAI #PersonalisedMedicine #PrecisionMedicine
-
The Attritio AI team built a drug program attrition risk assessment tool in less than 24 hours at a hackathon. The team used the Elicit API to retrieve structured evidence from the scientific literature, cite the underlying papers, and make the reasoning behind each conclusion easier to inspect. This is exactly why we opened Elicit's research capabilities to developers: so teams can build evidence-grounded research workflows directly into the tools and agents they are create. Learn more at https://lnkd.in/gN3Cf7Tk. Congratulations to the entire Attritio AI team, and thanks for the shoutout!
I had a fun and interesting time in London a couple of weeks ago at the "Building an AI Scientist" Hackathon (https://luma.com/yw0c3upd) co-hosted by Ternary Therapeutics, future.bio, pluto house, and Anthropic. Our awesome team built #Attritio, an app that estimates the chance a drug program fails in Phase 2 or 3. This was created in less than 24 hours, when only a few years ago could have cost six or seven figures to develop. It’s crazy to think what it will be possible to build in the next few years. Input a disease, target, and drug. It searches biomedical databases, including AMASS and Elicit, and returns a structured risk read-out: overall attrition risk (%), likely failure mode (safety- vs efficacy-driven), confidence level, the main reasons, and recommended de-risking experiments. The APIs from Elicit and Amass let us pull structured evidence from academic literature and other datasets, so we can cite real papers or data and understand the thinking that backed up the conclusions. Excited to see how AI can be leveraged to help de-risk drug hunting , drug repurposing, and help reduce R&D costs and risk. Thanks to the rest of the team, Cécile Soudé, Kiran Gathani, Oishi Deb, and Song Cao, and to Elicit and Amass for the API credits. You can try out the app here (until the credits run out 😅 ), maybe you will discover the next blockbuster drug: https://lnkd.in/dxQ-THcB #AI #DrugDiscovery #ClinicalTrials #drugrepurposing
-
Congratulations to the Apoptosis AI team on winning the "Building an AI Scientist" hackathon! The team built a multi-agent system for target validation and early discovery, using the Elicit API to ground its outputs in retrieved evidence. This is exactly the kind of workflow we built the Elicit API to support. Learn more at elicit.com/developer. We're excited to see what they build next. Thanks for the shoutout Janik Ludwig and team!
We won future.bio x Anthropic's "Building an AI Scientist" hackathon in London with Apoptosis AI -- point it at a target and a disease, and it comes back with a grounded GO/NO-GO/MAYBE, plus the highest-impact experiments to turn the maybe into a decision. Instead of asking an LLM to guess, Apoptosis runs the workflow a target team runs: separate agents pull genetics, biology, safety, chemistry, prior trials, patents -- and propose high-impact experiments on top of that, all grounded in evidence they actually retrieved. Elicit did a lot of the heavy lifting on the hardest part to get right: grounding. We wired it into several of our agents, each firing off its own literature searches. Elicit made us faster and more confident thanks to strong retrieval and verification tools: every citation the model makes gets cross-checked -- and we only touched a slice of the API (fast search). Big thanks to the Elicit team! Built by Janik Ludwig, Jan Boltersdorf, Gopalkrishna Purohit, Axay S., Youssef Abdalla, and Dmitry Kalupin. If you work in target validation or early discovery, we'd love to show Apoptosis to you -- and to hear where it falls apart: https://lnkd.in/em865GT7
-
The Elicit API and MCP server are now generally available with upgraded search and a new systematic review API. AI agents are becoming part of how scientific research gets done. But reliable research agents need access to grounded evidence. The API and MCP server bring Elicit’s trusted capabilities to your agents and workflows. Specifics of the end points now available: - Search API: This API allows you to search 138 million academic papers and 545K clinical trials using either semantic or keyword search. Our API had the best search results (paper recall) when compared against four of the most popular search tools on the BioASQ benchmark. - Reports API: This allows you to submit a research question and generate the answer as a fully cited report based on evidence from hundreds of papers or clinical trials. - Systematic review API: The most powerful of the three APIs is the Systematic Review API. This API lets you run a full systematic review end-to-end while giving you control over every step across search strategy, screening criteria, data extraction parameters, and reporting. All three also work over MCP, allowing you to connect Elicit as a Claude connector or ChatGPT plugin through their directories or to other MCP-compatible clients. Watch the short demo to see how it works.