The record your robots write to. Built for agents to read.

Robots and scientists write to one model, each at the detail they need.

A robot's work, in one record A robot arm sets a vial in a rack. Three threads carry its work side by side into one record, datin. One straight line runs on to datin's agent, the coral arrow in an app window, which reads the record and answers. The flow runs both ways: the robot's work travels into the record and on to the agent, and the record sends the robot its next step. Robots datin Agent

Robots produce more data, not more knowledge

I once passed a lab with more than $3M of equipment that spent most of its life idle. Hardware without a brain.

Nick Garabedian, Founder & CEO

Challenges

  • Every new workflow changes the schema

  • Agents need context

  • The in-house graph nobody maintains

You can build a graph on your own. Maintaining it is the job.

Many labs build a graph that works for a prototype. The hard part comes years later: keeping it full, migrating it, and connecting data reliably.

For any build, ask: who maintains it after the project ends?

Where in-house builds break →

What datin does in a robotic lab

  • People and robots in one model

    Robots and people record into one workflow, each at the detail they need.

  • New workflows without schema changes

    Add new workflows without reconfiguring the database.

  • Lineage on every run

    Every step is its own record: when it ran, where, who or what did it, what went in and what came out.

  • Samples, equipment and projects on record

    Samples, equipment and projects change over time. datin tracks all three side by side and links them on every run.

  • datin's agent

    Reads the record in the app and cites its sources.

  • Write via SDK

    Write runs to the record from your code.

  • Agent access

    Agents query the record through a read-only API.

  • Fewer manual steps to record

  • Equipment versions over time

Simulations and experiments in one record.

datin records each simulation like any other experiment, with its inputs, method and code version.Simulated and measured results meet on the same sample.Both inform your next automated run.

Bring your own simulation: DFT, molecular dynamics, FEA, CFD, CALPHAD, phase-field, reaction kinetics.

Simulations and experiments in one record Diagram: a simulation and a robot's experiment write to the same sample in one datin record. Their results meet there and inform the next automated run. datin Sample Simulation Next run Experiment

Your agents' memory for the physical world.

Coding agents work because repositories have structure.Experiments need the same: a machine-readable record with full lineage.

In 2016, the FAIR principles already asked for data that "computational agents" can find and reuse. The agents have arrived.

Scientific Data, 2016

The W.E.A.V.E. framework

W.E.A.V.E.: autonomy is woven, not installed Five layers woven into one record. Warp, your lab's context of samples, materials, equipment and projects. Embody, workflows as digital twins for people and robots. Automate, recording with fewer manual steps. Vision, agents on the record. Evolve, the learning loop. Evolve the learning loop Vision agents on the record Automate recording with fewer manual steps Embody workflows as digital twins, for people and robots Warp your lab's context: samples, materials, equipment, projects

Results and decisions, one graph

datin's graph tracks decisions too. Gates read each result in its own terms, and every decision logs its inputs, options and who made the call.

For your platform engineer

Bring one orchestrator workflow and see how it enters the record.Host on EU or US cloud, or your servers.Stored and exportable as RDF for full FAIR data.

  • EU or US cloud
  • On your servers
  • No outbound calls

How a robotic lab starts

No robots yet? Start here.

Start with your team.datin records each experiment your scientists run today, creating the memory for your future robots and agents.

Start with test labs →

  • Scientists, today
  • Robots, when you add them
  • Agents, when you add them

Show us one workflow.

  1. 15-minute call
  2. Scoping
  3. One workflow first, fixed fee

Deployed on-premises or in the cloud. Your data stays yours, and confidential.

Your contact

Dr. Nick Garabedian

Dr. Nick Garabedian

Founder & CEO, datin GmbH

Connect on LinkedInBook 15 min