Tuva AI

Agentic research collaborators for every human scientist

[Tuva AI @TuvaAI]

Take a look at our website: https://t.co/nGE3bcgTDp

[Tuva AI @TuvaAI]

Agents can help you do more science, but you still need to understand what they’re doing to direct them and communicate results to your colleagues. That’s why we built Rao’s new Reports feature.

Launch Rao to work on a problem or assemble findings from a conversation and report back to you in a concise, visual deck. Iterate on ideas by asking Rao to run new analyses and add slides. When it comes time to read out to others, assemble decks for talks or meetings in minutes, not hours.

Reach out to try it!

[Tuva AI @TuvaAI]

Check out matplotgym - our new benchmark for AI scientific figure generation!

[Vamsi Varanasi @feynmanpoint]

1/ A scientific paper starts and ends with its figures. A figure compresses a scientific claim and the evidence for it, and making the right figure is evidence that you understand the science you’ve done.

So: given a paper’s raw data, can a model create a figure that accurately represents a claim?

Introducing matplotgym, our new benchmark on scientific figure generation from @TuvaAI.

[Tuva AI @TuvaAI]

Our website: https://t.co/nGE3bchrsX

[Tuva AI @TuvaAI]

Introducing Tuva Reports: a new way to discover more, faster, by turning every scientist into a PI managing a team of agents.

Launch agents on long horizon tasks and have them report back to you in concise, understandable slides. All of the agents' work is backed by detailed traces and Rao's research memory so you can double check and dig into every result. Build on ideas by having agents add to decks with new results, all under your supervision.

Use Tuva Reports to have agents work on threads in parallel to you and generate precise decks for talks in minutes. Reach out to learn more!

[Tuva AI @TuvaAI]

Website: https://t.co/nGE3bcgTDp

[Tuva AI @TuvaAI]

Tuva's new subagent feature lets you launch subagents right from chat, so you can explore new ideas without losing your main train of thought.

DM us or check out the link below to try it!

[Tuva AI @TuvaAI]

At Tuva, we think of scientific research like a tree diagram. You start at layer 0, and from there you have n possible edges leading to different nodes on layer 1.

We think AI is most useful for surfacing more of those edges - helping you see all the directions your research could go before you pick one.

That's why we're building proactive agents that find those edges for you, based on where you are in your project, without you having to ask.

[Tuva AI @TuvaAI]

Our website: https://t.co/nGE3bcgTDp

[Tuva AI @TuvaAI]

An AI researcher at Princeton uses Rao to explore new research directions! He prefers Claude as a sounding board for ideas, but wants Codex to do preliminary analysis on his topics of interest. Rao lets him orchestrate agents across both.

Reach out if Rao could be useful for you! DM us or check out our website linked below.

[Tuva AI @TuvaAI]

Website: https://t.co/c3IcpBZIzu

[Tuva AI @TuvaAI]

The GPT-6 release might have you toggling between Claude Code & Codex - Rao’s new daily log makes it easier to keep track of the work you’ve done across both. The daily log automatically catalogs progress across Claude Code, Codex, and Rao.

As always, reach out if Rao could be useful for your work! DM us or check out our website linked below.

[Tuva AI @TuvaAI]

Our website: https://t.co/nGE3bcgTDp

[Tuva AI @TuvaAI]

A biophysics PhD student uses Rao to connect his conversations across Claude Code, Codex, and Rao itself. By centralizing context, he uses Rao to look across his entire workspace and recognize open threads.

If Rao could be useful for your research, we’d love to talk! DM us or visit our website linked below.

[Tuva AI @TuvaAI]

Register here: https://t.co/EZQgrgJquS

[Tuva AI @TuvaAI]

Join us on Sep 14 to chat about how we can use AI to do more science, faster!

Food & drinks will be provided, along with the company of a room full of fellow AIxBiotech builders. Link in the comments below!

[Tuva AI @TuvaAI]

https://t.co/c3IcpBZIzu

[Tuva AI @TuvaAI]

Excited to announce the the new Rao desktop app! Here, you can coordinate agents that can work across the Rao interface, Claude Code, and Codex and act based on where your project actually is.

One new agent runs proactive literature reviews, referencing your project context to find papers most relevant to your recent progress and ideas. It’s hard to know when inspiration will be most helpful and it’s often tucked away in years-old papers, so we’ve designed this Rao agent to inspiration-hunt for you.

Visit our website or DM us to try Rao! Link in the comments.

[James Sinka - Boston & SF @jamessinka]

Thank you to our partners who make this possible

@AnthropicAI
@nvidia
@PrimeIntellect
@answerthisio
@AppliedSciAI
@muni_bio
@RowanSci
@TuvaAI

[Vamsi Varanasi @feynmanpoint]

Today, we’re launching @TuvaAI to build a human-centric future for agentic science.

Humans collaborating with AI can learn more than either working alone: humans guiding the research, equipped with agents that can test ideas quickly and autonomously, will deliver the next wave of scientific discovery. But science is all about the details. Neither Einstein nor AGI can drive your project forward without intimately understanding your growing, interrelated web of hypotheses, intermediate results, working threads, failed experiments, datasets, notes, and water cooler conversations.

Our first product, Rao, was born out of my frustration having to reteach LLMs these constantly changing details for every single task in my own research. Rao maintains a persistent, dynamic research memory that learns and remembers project context, history, and threads. This research memory then links into our scientific agents as well as other AI software you may already be using, such as Claude or Codex. With Rao, every scientist turns into a PI managing a proactive, agentic research team that automatically stays on the same page.

Rao is live in VS Code and by CLI, and can help with any scientific task that can be done in a computer. Our early users include computational biologists, theoretical neuroscientists, chip designers, and polymer physicists who use Rao daily to ideate, do math, analyze data, and write papers and grants.

Now, we’re excited to announce Rao in closed beta. If your science mostly happens in a computer—theory, computational modeling, experimental data analysis, scientific writing—Rao can help you discover more, faster. Sign up for our waitlist, and if you’re a fit, we’ll work with you to tailor new, contextual agents that augment the way you do science. Link in the comments.

More personally: I’ve been doing scientific research since I was a kid. Scientists do so much for the world, choosing a painstaking career for the joy of discovery, developing technologies we use every day along the way. And yet, the tools scientists use are often cumbersome and woefully out of date. For me, Tuva is much about empowering the people of science with products they love as it is about the discoveries we’ll help them make. And I’m lucky to work with some wonderful humans along the way—@SamsaraDurvasu1 @anuvellore @KimchiOfer among many others. If our mission resonates, please reach out.