Sous · 001 Start now

Your AI lab assistant

Collect FAIR data without hassle. Leverage your experimental history to design better experiments and interpret results.

Sous · 001 Start now

A connected workflow for every experiment

Turn protocols into structured, runnable workflows.

Sousy points to a structured protocol card.

Upload raw protocol files, start from a standard library, and keep every change versioned so the exact method behind each experiment stays clear.

Golden Gate Assembly shown as a structured protocol with inputs, assembly steps, a validation decision, and pass or retry paths.

Everything stays connected

Built for the whole experimental workflow.

01 / Methods

Structured protocols

Import and organise reusable methods as clear, runnable protocols.

02 / Methods

Version history

Know the exact protocol version behind every experimental run.

03 / Methods

Visual workflows

Express steps, decisions and repeats as a workflow the whole lab can follow.

04 / At the bench

Guided runs

Set run parameters and move through the protocol one action at a time.

05 / At the bench

Bench records

Keep notes, deviations and result files beside the step that produced them.

06 / Experiments

Experiment notebook

Bring hypotheses, plans, runs, results and conclusions into one record.

07 / Experiments

Project context

See how pilots, validations and follow-up experiments connect in one graph.

08 / Lab operations

Sample provenance

Trace created samples back to the protocol run that produced them.

09 / Lab operations

Inventory & locations

Search samples across nested lab locations, including addressable plates.

Your sous chef at the bench

Ready to spend less time on lab paperwork?

Bring protocols, experiment records and research context into one connected workflow.


Frequently asked questions

Using Sous

Who is it for?

Wet-lab teams that want an easy, consistent way to record their methods and results, and who want powerful, scientifically grounded AI to refine their experimental plans and conclusions.

What makes it different from a static SOP or PDF?

Unlike a static document, Sous keeps every experiment linked to the exact protocol version used and records changes made at the bench.

Can I still take notes by hand at the bench?

Yes. Sous can parse handwritten notes and fold them into the experiment record automatically, so you are not forced to type everything in real time. Write how you write; the structure gets added for you.

I hate AI. Can I use Sous without using AI features?

Sure. AI features are only ever triggered if you call for them.

Plans and AI

Will Sous be free for academia?

Yes. Sous will always have a free plan for academic labs. We may charge for optional bonus features in the future, but the core academic workflow will stay free.

Where do the suggestions come from?

Sous draws on two sources: published scientific literature and your own experimental history. When you are planning experiments or interpreting a result, Sous analyses both to provide suggestions in context.

Privacy and security

Is my unpublished data safe?

Your data is not used to train models. Raw data and unpublished results stay private to your team until you choose otherwise. If you choose to use LLM-powered suggestions, then experimental plans and conclusions (but no raw data) may be shared with third-party LLM providers under terms that do not allow them to use that information.

How does Sous protect lab data?

Access is tied to authenticated accounts and organisation membership. Experiment records, protocols, runs, notes, and result files are stored behind row-level security policies, so users can only read or change data for organisations they belong to. Admin-only actions, such as deleting some shared records, are checked separately.

How is data isolated between teams?

Each core record carries an organisation ID, and database policies check membership before returning or accepting changes. PowerSync uses the same organisation membership boundary when deciding what data to sync to a device, so a user only receives the local working set for teams they can access.

Ask another question