Noemsera hypothesis-native · ai for science
AI for Science · biomedical HYPOTHESIS-NATIVE AGENT PLATFORM

Where biomedical data
becomes scientific reasoning.

Noemsera turns medical signals, literature, and experimental context into testable hypotheses — and into discovery loops that remember the project.
See the substrate
noema · object of thought sera · biomedical signal self-hostable
§ 01 · the name
Etymology

The name names the work.

A claim that can't be traced from sample to reasoning is a guess. The name is a promise not to make any.

noema
/ˈnoʊ.ə.mə/ · phenomenology · Husserl
The object of thought — what is held in consciousness, distinct from the thinking. Examinable. Reasoned-about.
sera
/ˈsɪr.ə/ · plural of serum
The biomedical signal — serum, imaging, EHR, sequence. The raw material of clinical reasoning.
noema + sera = medical evidence as objects both a scientist and an agent can reason about.
§ 02 · the substrate
From signal to hypothesis

A substrate, not a chat window.

Biomedical research is data-rich and hypothesis-constrained. Noemsera turns evidence into reasoning-ready objects. Autonomous agents generate hypotheses, evaluate mechanism, plan validation.

§ 03 · the loop
Discovery loop

From observation to experiment — and back.

Seven capabilities close the circuit. Not a passive assistant — an active reasoning system that amplifies the loop.

01

Evidence understanding

Ingest biomedical evidence — literature, omics, EHR, imaging — as structured context, not unstructured text.

02

Scientific object modeling

Organize data into reasoning-ready primitives: diseases, pathways, targets, cohorts, endpoints.

03

Mechanism-aware reasoning

Reason over mechanism, pathway connectivity, cell–cell communication, and causal chains — not keywords.

04

Hypothesis generation

Produce ranked, falsifiable hypotheses — with the evidence that supports and the evidence that refutes.

05

Evidence critique

Identify supporting and contradicting evidence, quantify uncertainty, surface the missing experiments.

06

Validation planning

Propose concrete next steps: experiments, cohort analyses, multi-omics validation paths, replication designs.

07

Discovery-loop memory

Accumulate project memory so the agent evolves with the study, not against it.

The loop

The cognitive circuit closes: observation → mechanism → hypothesis → experiment → observation.

§ 04 · the workbench
A repo, not a chat history

Sessions, branches, phase gates.

A project is a git workspace, a session a branch, a phase a signed tag. The loop is the repo.

fig.02 · project · sessions · phase gates cohort-upstaging
cohort-cleaning
mechanism-hypothesis
+ new session
project/main
session/a
session/b
P0
Sandbox
done
P1
Cohort
done
P2
Variables
done
P3
Mechanism map
tree
P4
Models
P5
Critique
P6
Validation
P7
Manuscript
two modes
Research runs an SOP from cohort to manuscript. Discovery turns raw data into a connector other projects can mount.
connectors
Bundled with mimic-iv eicu-crd lung-cohort. Your own lands in ~/.noemsera/connectors/.
deploy
On your laptop, or your lab's box behind Cloudflare Access and netbird. Patient data stays on your network.