The AI Scientist: Automating Scientific Discovery with AI

« The AI Scientist, from Sakana, is presented as a system that automates scientific research end to end rather than assisting with one part of it. That is an ambitious claim, and it belongs in the research-frontier category — interesting to follow rather than something to depend on for work that matters. »

Like this tool? We can build a custom agent that does this and connects to your own systems.

Build a Custom Agent

freemium670

Similar AI Tools

Materiom

Papers GPT

Cradle

AlphaGeometry by Google

WeatherNext 2

#Research & Science

The AI Scientist: Automating Scientific Discovery with AI at a glance

Category: Research & Science
Category rank: #1 of 43
Pricing: Freemium
Popularity: 239 upvotes
Last updated: Aug 27, 2026

API

Not available

Free tier

Yes

Key features

Aims at automating research end to end rather than assisting one step.
Comes from Sakana as a research system rather than a commercial product.
Sits at the frontier of what automated research can do.
A free tier makes it possible to explore the approach.

Pros

Genuinely ambitious scope rather than an incremental research assistant.
A free tier makes it possible to see what it actually does.

Cons

Automated research output needs the same scrutiny as any other — arguably more, since nobody chose the method.
Frontier systems are for exploring rather than depending on.
Publishing automated research without disclosure raises real integrity questions.

Use cases

Exploring what end-to-end research automation currently produces.Following the frontier of automated scientific work.Prototyping an idea that a researcher then takes forward properly.

Industries

ResearchEducationTechnology

FAQ

It is a system from Sakana presented as the first to automate scientific research.

There is a free tier. Check the current terms for what is included.

Research integrity requirements apply regardless of what produced the work. Disclosure expectations around automated research are still developing, so check the norms in your field.

It needs at least as much scrutiny as human research, arguably more, since no researcher chose the method or judged whether the question was framed well.