TeleonomyResearch infrastructure

The operating system for autonomous science.

One place to plan research, run experiments, and analyze data — so the why is captured as you go.

Instruments: Connected
Compute: Managed
Agents: Governed
Record: Automatic

The old problemTacit knowledge

Most science is not written down. It is remembered.

A paper tells you what was done. It rarely tells you what made it work. The working method lives with the people who run it — the reagent that behaves and the one that doesn't, the change in a sample that tells you something has gone wrong, the instrument that needs a minute longer than the manual says.

We have called this tacit knowledge for seventy years, and lost it to staff turnover for longer than that.

Every failed replication is a record that was never complete enough to follow.

Why nowAutomation scales it

Automation multiplies an incomplete record.

Results arrive faster than anyone can reconstruct how they were produced, and nothing in the run records why it was set up that way.

In regulated science, a result you cannot trace is a result you cannot file.

The systemHow the record happens

Teleonomy directs the work and writes the record.

A run is directed from one place. Scientists, agents, instruments and compute all work through the same system, with each action governed by your policies.

The record is created as the work progresses. It captures what was done and how, and asks for the reasoning when the decision is made, rather than when the report is due. Any result can be traced back to the data, the method and the decisions that produced it.

An experiment's protocol: numbered steps with their parameters, a hazard note, and the expected result for each one.

A table of laboratory devices with model, serial, location and calibration date. One hazard flag is open, showing a maintenance note about priming the flow cells before capture.

A dialog for adding an analysis to a run: a searchable list of code kits, filtered to those compatible with this run's data.

An assistant proposing to update an experiment's results, stopped by a permission prompt naming the fields it would change, with Allow and Deny.

Fig. 1 | (a) A protocol and its expected results. (b) Instruments, with calibration dates and handling notes. (c) An analysis chosen for a run. (d) An agent stopped for permission before writing.

The record is a by-product of the work, not a task that follows it.

Where this goesToward autonomous labs

Today, and the road ahead.

Today: Run experiments with versioned protocols and data
Track materials, samples, devices and software
Analyze data with compute provisioned for each run
Connect instruments for automatic data collection
Set governance policies, enforced on every action
Manage teams and access in shared workspaces
Next: Operate connected instruments from the same system
Share findings with the evidence behind them
Ahead: Close the research loop: agents design, run, interpret and decide what follows

Get startedAccess by request

The future of research starts here.

And it builds on everything you have already done.

Request access hello@teleonomy.ai