Undermind
Deep AI literature search that trades speed for near-total recall
Overview
Undermind is an AI agent that runs adaptive, multi-round searches of academic literature, reading full papers to approach exhaustive coverage of a topic. Deep searches take 8–10 minutes and return annotated, cited reports; the free plan allows a few searches a month.
Pricing
Pricing shown for reference only. These figures reflect RECATOOLS research as of 11 Jul 2026 and may be out of date or incomplete. This is not financial or purchasing advice — always confirm the current price on the provider’s official website before making any decision.
- Strong AI models for chat/search
- Report generation
- Shared project collaboration
- Standard rate limits
- Latest, most powerful AI models
- Deepest analysis of full texts
- 10x higher usage limits
- Unlimited projects & files
- All Pro features
- Team member management
- Priority support
- Centralized billing
- Increased compute
- Sitewide SSO
- Admin dashboard
- Custom SLA & security review
Use cases
ASEAN Perspective
Undermind in Southeast Asia
ASEAN-region availability and pricing notes coming soon. Drop the editorial team a note via /contact/ if you can supply local context (Singapore/Malaysia/Indonesia/Thailand/Vietnam).
Undermind, a YC-backed startup from two MIT physics PhDs, runs successive adaptive searches over the scientific literature instead of one-shot keyword retrieval, reading full texts and chasing citation trails to approach exhaustive coverage. Each deep search takes eight to ten minutes by design — you trade speed for recall, and for defensible literature reviews that trade is usually worth it. The free plan allows only a few searches a month; Pro runs about $16–20/month. The company's 98% precision claim is its own benchmark, not independently audited, and a four-person team carries continuity risk. It's specialised for literature discovery, not general Q&A. English-first, globally available, usable from ASEAN.
What people say
Eight to ten minutes per search. That's both the pitch and the price of Undermind: instead of returning a ranked list in half a second, the agent runs successive rounds of semantic, keyword and citation searches, reads full papers rather than abstracts, follows citation trails, and comes back with an annotated report that tries to catch nearly everything relevant. For work where missing a paper is expensive — grant applications, PhD onboarding, tech due diligence — trading speed for recall is exactly right.
The company is tiny: founded in 2023 by two MIT physics PhDs (Thomas Hartke and Joshua Ramette), Y Combinator-backed with about $500K raised, four employees, and roughly $440K in 2025 revenue per GetLatka. Pricing is modest and reported inconsistently across review sites — the free plan allows a handful of deep searches a month (3 to 5 depending on the source), and Pro sits around $16–20/month. Team and enterprise plans are quote-only.
Two caveats. The 98% precision figure that shows up in every review is Undermind's own benchmark, not an independent audit. And a four-person startup is a continuity risk if a lab standardises its workflow on it — export what you find. It's also strictly a literature tool: English-first, academic databases only, no general-purpose chat.
Reviewers keep landing in the same place: niche, slow on purpose, and genuinely better at exhaustive discovery than one-shot search engines or a chatbot's confident summary. If your job involves defensible literature coverage, it earns a serious trial.
Summary of public user & expert reviews, compiled by RECATOOLS.
Notable facts
- Undermind can autonomously read and synthesise 200+ papers in the time it would take a researcher to read 10, producing a literature summary that covers the field comprehensively.
- The agentic architecture follows citation chains — if a relevant paper cites an even more relevant paper, Undermind will find and read it without being explicitly directed.
- Undermind was built specifically because the founders noticed that even excellent researchers spend 30-40% of their time on literature search and synthesis rather than novel work.
Frequently asked questions
About this listing
This entry was compiled from publicly available data including Undermind's official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with Undermind unless explicitly stated.
Third-party AI tools update their pricing, features, availability, and policies frequently. Information here may be outdated by the time you read this — we make reasonable efforts to keep listings current, but cannot guarantee absolute accuracy.
For the latest details, please refer to Undermind directly →
Spotted something out of date? Suggest an update →
More in Research & Data
