Undermind
Undermind is a standalone vertical AI agent from Undermind. A scientific-literature agent that searches thousands of papers, reads them and writes up what the evidence says. Nearfi's Agent Index rates its autonomy 4/5 (multi-step on its own): it plans and runs an adaptive literature search on its own and returns an evidence-reviewed report; it is not 5 because it runs only when asked, and not 3 because you don't approve each step. Status: live; 0 incident(s) logged.
By Undermind · launched unknown · last checked 2026-11-04
Why 4: it plans and runs an adaptive literature search on its own and returns an evidence-reviewed report; it is not 5 because it runs only when asked, and not 3 because you don't approve each step.
What it does. You describe a research question and Undermind runs a deep search, reviews thousands of papers and adds the relevant ones to a workspace with an evidence-quality review. The maker says a search stops by itself after about 2.5 minutes on average. It fetches open-access PDFs, lets you upload others, and can produce cited writeups; it can also connect to Claude, ChatGPT and other agents.
Where it falls short. It covers academic literature rather than the open web. Its recall and precision figures come from the company's own 23-query benchmark and haven't been independently confirmed; the pricing page gives no numeric usage limits.
What it can do
Incidents
- None logged.
Closest fiction
Contact — A machine that reads everything so a person can ask one good question.
Related research agents
- ChatGPT deep research OpenAI · autonomy 4/5
- Consensus Deep Search Consensus · autonomy 4/5
- Elicit Research Agent Elicit · autonomy 4/5
- Gemini Deep Research Google · autonomy 4/5
- Perplexity Deep Research Perplexity · autonomy 4/5
History
- 2026-11-04 Added to the index.
Sources
← All agents in the index · Autonomy is rated on what the product does today, on a 1–5 scale. Every fact links to a source.