Skip to content
CITATION LINKED Executive decision support
Literature and target-discovery agent

Cohort

Searches across the literature and your internal assay data together, and every claim it makes carries the paper it came from.

150 tokens/s Throughput
99.0% Accuracy
2.1s Median latency
1.9M Calls per month

Retrieves across published literature, trial registries and internal experiment records; distinguishes replicated findings from single-study results; states when the evidence is thin.

Scope

Retrieves across published literature, trial registries and internal experiment records; distinguishes replicated findings from single-study results; states when the evidence is thin.

Capabilities

  • Cross-corpus retrieval
  • Replication status per finding
  • Trial registry linkage
  • Internal assay integration
  • Every claim cited

Runs on

Claude 3.7 Sonnet Gemini 2.5 Pro
Calling it from the SDK
const research = new NexusAgent({ role: 'research', corpora: ['pubmed', 'internal://assays'] });

const brief = await research.investigate({
  question: 'Evidence for target X in fibrotic disease',
  requireCitations: true
});

Recorded exchanges

Published examples, reproduced verbatim. This page does not call a model.

How strong is the evidence, really

Asked
Summarise the evidence for this target in idiopathic pulmonary fibrosis and tell me how much of it is replicated.
Returned
Fourteen relevant publications since 2019, plus three internal assay runs.

Replicated: the expression increase in fibrotic tissue appears in four independent cohorts across three groups (citations attached), with consistent direction and effect size within a factor of two. This is solid.

Single-source: the proposed mechanism rests on one 2022 paper. It has 90 citations, but I could find no independent replication of the knockdown result — the citing work largely repeats the claim rather than testing it. Treat the mechanism as a hypothesis.

Contradicted: two of the three internal assay runs did not reproduce the published potency, at roughly a fifth of the reported effect. Worth understanding before this goes further.

I am summarising evidence, not recommending a programme decision.
Drug discovery Literature Evidence
All agents