Summary
Elicit is a hosted AI research agent for science, searching over 138 million papers plus clinical trials, patents and regulatory databases. It screens and extracts evidence into cited reports, tables and systematic reviews, with every claim traced back to a source sentence.
- Best for
- Academic researchers, R&D and policy teams, and pharma strategy groups running literature reviews or formal PRISMA-2020 systematic reviews who need cited, structured evidence pulled from published research.
- Not for
- Buyers wanting a general web-search answer engine (see Perplexity), a general-purpose autonomous agent that automates open-web tasks and deliverables beyond published literature (see Manus), or a cheaper literature-only discovery agent that does not run the formal PRISMA 2020 systematic review workflow (see Undermind). Elicit is scoped to the academic and scientific literature, not the open web.
- From
- Freethen $75 monthly · $49 billed annually
- Free tier
- YesBasic plan: limited monthly usage for Research Agent and Research Reports, unlimited paper search across 138 million papers, unlimited summaries and chat with papers, Zotero import. Systematic Review, Extract Data and API access are excluded (API: "Basic | No access"). When usage runs out, only quick tools like Find Papers and Chat with Papers remain until the monthly reset.
What it does
- Done-for-you service
- NoSelf-serve software: Elicit states it is usable immediately with no implementation project.
Elicit is immediately available for your researchers to use. There is no lengthy implementation
elicit.com/industries/pharma
How much it does unattended
- Runs autonomously
- PartlyPlans and executes multi-step research unattended, but asks clarifying questions before it starts and expects you to review the output.
The Research Agent breaks your prompt into a systematic plan, then executes it step by step.
support.elicit.com/en/articles/14756886… - Agent permissions
- YesSkill creation is gated on explicit approval, and the agent asks scoping questions before it spends usage.
Nothing is saved until you approve it, and you can create, revise, or cancel at that point.
support.elicit.com/en/articles/16647708…
Not established: Multi-agent
Which models it runs on
- Claude
- PartlyClaude Opus has been shipped in the product, but announced only for paid tiers, not the free Basic plan.
Opus 4.5 is now live to Elicit users on Pro, Teams, and Enterprise.
elicit.com/blog/evaluating-claude-opus-… - GPT
- PartlyElicit names OpenAI as a third-party model provider it sends data to, but never states which GPT model powers the product.
We explicitly do not train on Enterprise user data, and have agreements with third-party providers (like OpenAI) that prevent them from doing so…
elicit.com/blog/elicit-achieves-soc-2 - Model choice
- NoElicit selects and ensembles models itself; the only user control is an effort slider from Fastest to Smartest, not a model picker.
We use strategies like process supervision, prompt engineering, ensembling multiple models, double-checking our results with custom models and…
support.elicit.com/en/articles/14757979…
Not established: Open models
Where you use it
- In your editor
- NoElicit enumerates its surfaces as the web app and the API; the only installable client is a Chrome/Edge browser extension.
The Research Agent is available through elicit.com or via API to bring into your workflows.
elicit.com/blog/introducing-elicit-rese… - On the command line
- PartlyA command-line tool ships as an integration example in Elicit's own API examples repository; no CLI is sold or documented as a product.
Working examples in curl, Python, and JavaScript, plus integrations (CLI tool, Slack bot, Claude Code skill):
docs.elicit.com - In your pipeline
- NoNo CI or pull-request surface; Elicit's own surface enumeration is the web app and the API, and the closest published case is a customer's own scheduled jobs.
The Research Agent is available through elicit.com or via API to bring into your workflows.
elicit.com/blog/introducing-elicit-rese… - In a browser
- YesHosted web app is the primary surface, with all workflows launched from the home page.
On Elicit's home page, you'll find several different options to start your research: Research Report, Systematic Review, Find Papers, Chat with…
support.elicit.com/en/articles/14757543…
Whose machine it runs on
- Self-hosted
- PartlyEnterprise offers custom deployments including single-tenancy; Elicit publishes no on-premises, VPC or air-gapped option.
Enterprise-level security and controls, including SSO & SAML, 2FA, user analytics, domain verification, and custom deployments (e.g. single-tenancy)
elicit.com/pricing - Open source
- PartlyOnly Elicit's API examples repository is public and MIT licensed; the platform itself is a closed hosted service.
Working examples in curl, Python, and JavaScript, plus integrations (CLI tool, Slack bot, Claude Code skill):
docs.elicit.com
What it costs to run
- How it meters
- YesPer-seat monthly subscription with one shared monthly usage pool; Pro and Scale can enable extra usage, billed in $100 increments against a limit you set.
You set a monthly spend limit; once you reach it, extra usage stops until the next period. You're never charged beyond the limit you set for a given…
support.elicit.com/en/articles/15646622… - Free tier
- YesBasic plan is free with limited agent and report usage but unlimited paper search; no sales call required.
Limited usage for Research Agent and Research Reports
elicit.com/pricing - API access
- PartlyAPI is included from Pro upward and absent on Basic and Plus; Enterprise is billed as unlimited API access.
The Elicit API is available on the Pro, Scale, and Enterprise plans.
support.elicit.com/en/articles/14757400… - MCP server
- PartlyA hosted MCP server exposes the full API over OAuth 2.0 to any MCP client, gated to Pro and above.
All API functionality is also available via MCP (Model Context Protocol) server, enabling use from Claude Desktop, Claude Code, and other…
docs.elicit.com - Bring your own key
- PartlyYou bring your own documents and your institution's publisher subscriptions; the vendor documents no way to supply your own model keys or inference.
Upload your own documents and integrate company subscriptions to access more full-text papers.
elicit.com/solutions/research-agent
Buying it for a team
- A company can buy it
- YesScale is self-serve for teams with centralised billing; institutions and enterprises can buy by invoice.
Centralized billing: All team members' subscriptions are billed to the Team Admin in a single purchase.
support.elicit.com/en/articles/14759330… - Seat model
- YesPriced per user per month; adding a seat immediately raises included usage and you need not fill it. No seat minimum or ceiling is published.
Adding seats will immediately increase your team's included usage and allow you to invite an additional user.
support.elicit.com/en/articles/14759330… - Pooled budget
- YesA Scale team's included usage is the sum of its seats and drawn from one shared pool.
A team's included usage is the sum of its seats, so a 4-member team has 4x an individual's usage, and the whole team draws from one pool.
support.elicit.com/en/articles/14759330… - Admin controls
- PartlyAdmin dashboard covers seats, usage and immediate lockout of a member; no published controls over which features, data sources or agent actions users may use.
Team admins have access to an Admin Dashboard to invite or remove team members, view usage, and purchase additional seats.
support.elicit.com/en/articles/14759330… - Audit log
- PartlyTeam actions and artifact edits are logged with attribution and version history, and screening decisions are PRISMA-auditable; no exportable security or access log is published.
All actions and edits are logged, attributed to the team member who took that action, and kept in version history.
support.elicit.com/en/articles/16649941… - Single sign-on
- PartlySAML-based SSO with 2FA and domain verification, sold only on Enterprise; OIDC is never named.
We support SSO & SAML-based authentication systems to ensure seamless account management even at 10,000+ users.
elicit.com/industries/pharma
What happens to your code
- Opt out of training
- PartlyNo training on Enterprise data by default, with zero-retention agreements downstream; no self-serve opt-out is documented for Basic, Plus, Pro or Scale.
Enterprise customer data is fully redacted. We do not train on this data. Our third party model providers also have zero data retention policies and…
elicit.com/industries/pharma - Data residency
- YesUnited States. EU transfers are handled contractually through Standard Contractual Clauses, not by EU hosting; no regional hosting option is published.
Our servers are located in the United States.
elicit.com/operations/privacy - Getting out
- YesTables export to CSV, Excel, RIS or BIB and reports to PDF or Word; the library exports as .ris and the account is self-service deletable. Unused monthly usage does not roll over.
Tables can be exported in CSV and Excel format. Certain tables of sources can be exported as RIS or BIB files.
support.elicit.com/en/articles/14758189… - Certifications
- PartlySOC 2 Type II with a clean opinion, and GDPR handled through a DPA with Standard Contractual Clauses; ISO 27001 and HIPAA are not claimed anywhere readable.
Elicit's SOC 2 Type II report did not have any noted exceptions and was issued with a “clean” audit opinion.
elicit.com/industries/pharma
27 sourced claims on this page. Checked 2026-09-02. How we check.
Alternatives
Where we would send a reader instead, in that listing's own words.
Open-source autonomous research agent that gathers sources and writes cited long-form reports.
General-purpose AI agent that plans and executes tasks in its own cloud computer.
AI literature research agent that searches, reads and cites the scientific record.
Filed under
Sources and updates
The pages we read21
Every fact below comes from the vendor. Nothing here is independently corroborated yet.
- elicit.com/pricing5 fields
- support.elicit.com/en/articles/14759330-elicit-s-scale-plan-for-teams4 fields
- elicit.com/industries/pharma4 fields
- docs.elicit.com/3 fields
- elicit.com/blog/introducing-elicit-research-agent3 fields
- support.elicit.com/en/articles/14756886-elicit-s-research-agent2 fields
- trust.elicit.com/subprocessors2 fields
- elicit.com/1 field
- trust.elicit.com/1 field
- elicit.com/blog/elicit-achieves-soc-21 field
- elicit.com/blog/evaluating-claude-opus-4-51 field
- support.elicit.com/en/articles/14757400-elicit-s-api1 field
- support.elicit.com/en/articles/14757404-use-elicit-via-mcp-server1 field
- support.elicit.com/en/articles/14757543-getting-started-with-elicit-which-workflow-or-tool-should-i-use1 field
- support.elicit.com/en/articles/14757979-elicit-s-reliability-does-elicit-run-the-same-risk-of-hallucination-as-other-ai-tools1 field
- support.elicit.com/en/articles/14758189-export-your-data-from-elicit1 field
- support.elicit.com/en/articles/15646622-usage-limits-in-elicit1 field
- support.elicit.com/en/articles/16647708-skills-in-elicit1 field
- support.elicit.com/en/articles/16649941-collaborate-in-a-shared-research-agent-session1 field
- elicit.com/operations/privacy1 field
- elicit.com/solutions/research-agent1 field
Updates3
- Capabilities · Commercial terms · Free tier · Price
- Capabilities · Commercial terms · Free tier · Price
- Alternatives · Crunchbase link · LinkedIn link · One-liner · Summary
