BIOLOGICAL INTELLIGENCE SYSTEMS

Make biology
predictable.

terminalBio builds world models that turn biological complexity into interventions you can design, test, and learn from.

Explore the platform ↓
BIOLOGICAL STATE
INTERVENTION SPACE
OBSERVED
STATE 01
PREDICTED
OUTCOME 02
MODEL CONFIDENCE0.94
SCROLL TO EXPLORE ↓

THE THESIS 01

Biology is not a list of parts.
It is a living system.

Today’s models can recognize patterns in sequence or structure. The next frontier is understanding how molecular events propagate through cells, tissues, and patients—and using that understanding to make better choices.

We’re building the computational layer for that work.

MOLECULARCELLULARTISSUEPATIENT

THE PLATFORM 02

A closed loop between computation and experiment.

The Biological
World Model

A continuously learning representation of biological state, response, and uncertainty. It connects evidence across scales so every intervention starts with a better hypothesis.

Explore the system →
TERMINAL
BIO
01 / OBSERVE

Integrate multimodal evidence—from molecular assays to longitudinal clinical signals—into a coherent starting state.

THE SYSTEMS 03

Specialized models, one shared intelligence layer.

01

BindFM

Molecular interaction intelligence for finding the forces that make binding work.

MOLECULAR
02

XUZU

Generative aptamer design with a direct line from sequence to experimental validation.

GENERATIVE
03

RNJ / RNX

RNA structure intelligence for a more legible, designable RNA landscape.

RNA
04

TB Biocompute

Composable infrastructure to run, compare, and build on biological models.

INFRASTRUCTURE
01 / BINDFM

Molecular interaction intelligence.

BindFM represents proteins, nucleic acids, small molecules, and cofactors in one interaction-aware space—so teams can rank candidates and decide what to test next.

USE CASEBinder discovery & interface designLOOP ROLERanks candidates for validation

INTERVENTION ENGINE

What should
we change?

Search the intervention space, simulate consequences, and prioritize the designs that move a biological system toward a desired state.

01
Current state
02
Select intervention
03
Desired state

EXPERIMENT ENGINE

What should
we test next?

Design experiments for information gain—not just throughput. Every assay is selected to sharpen the world model.

WHERE IT APPLIES 04

From a molecular
question to a decision.

01

Binder & aptamer
discovery

Generate, rank, and validate new binders against difficult targets.

Target and binder interaction map
02

Resistance
prediction

Anticipate escape paths before they arrive in the clinic.

03

RNA
therapeutics

Reason over structure, delivery, and function together.

DNARNAFUNCTION
04

Programmable
biology

CRISPR, protein engineering, and synthetic biology with system-level feedback.

A MODEL THAT GROWS 05

Across every
scale of biology.

01MolecularInteractions & structure
02CellularState & response
03TissueContext & dynamics
04PatientTranslation & outcomes

COMPOUNDING ADVANTAGE 06

Every experiment
makes the next one
more intelligent.

Our moat is not a static dataset. It is a growing feedback system: proprietary experimental readouts, high-resolution metadata, and models trained to decide what to measure next.

COMPOUNDING EVIDENCE
01

Experimental readout

02

Model update

03

Sharper next experiment

RESEARCH & OPEN SCIENCE 07

Build the field
with us.

We publish, collaborate, and create tools that make biological intelligence more useful to the people advancing it.

TERMINALBIO 08

Build a more
legible biology.

We are scientists, engineers, and builders working at the frontier of biological intelligence.

Meet terminalBio →

START A PROGRAM

Bring us your
hardest question.

Talk to terminalBio ↗