Skip to content

CAR-T Intelligence Agent

live   Domain: clinical   Type: agent

RAG decision support for CAR-T; stores scFv/antibody sequences.

A narrated, captioned explainer. Decision support for a qualified clinician.

CAR-T Intelligence Agent — what it takes in, what it computes, what it returns

Illustrative. Decision support for a qualified clinician — never autonomous diagnosis or prescribing.

In plain terms

The CAR-T Intelligence Agent is a research companion for cell-therapy design. It gathers and compares the evidence behind a CAR-T construct — the antigen it targets, the binder (scFv/antibody) sequences that recognize it, and what the literature says has worked — and hands a clinician or scientist a grounded evidence brief. It reasons; it does not design a therapy on its own.

Why it matters

CAR-T can drive durable remissions, but designing and choosing one is an evidence-heavy, high-stakes exercise spread across scattered literature and sequence databases. Pulling that evidence into one comparative, cited view is where an intelligence layer earns its place.

For a patient: the scattered evidence behind a cell therapy pulled into one clear, cited brief for the team designing their treatment.

How it works

How the CAR-T agent reasons — target antigen to comparative evidence to design brief

Cross-collection evidence and comparative analysis for CAR-T design. Decision support. Illustrative.

  1. Frame the target — the target antigen and patient context.
  2. Retrieve across collections — cross-collection evidence, including stored scFv/antibody sequences.
  3. Compare — comparative analysis and deeper research across candidate constructs.
  4. Ground the answer — a cited evidence brief; it refuses to fabricate where evidence is thin.

What goes in, what comes out

  • In: a query and the patient context.
  • Out: a grounded, cited CAR-T design evidence brief.

Where it fits

Where the CAR-T agent sits — reasoning with structural biology and single-cell

It draws on structural biology and single-cell context; anchor of the CAR-T demonstration. Illustrative.

It reasons alongside the structural-biology and single-cell capabilities and anchors the oncology / CAR-T design demonstration.

Honest limits

  • Decision support, never an autonomous decision. It supports a qualified clinician or scientist; it does not make or design treatment decisions on its own.
  • Grounded, and honest when it can't be. As a retrieval-augmented service it needs a populated vector database and an LLM API key at runtime, returning an honest degraded response (HTTP 503) rather than inventing content when they're absent.
  • Design frontier is separate. De-novo binder design (Chai-2) is a distinct, gated frontier capability — this agent reasons over evidence, it does not generate binders.

Interface

  • Endpoint: localhost:8521 · Invoke path: /
  • Serving: native · GPU: no · Cost class: —

Inputs

Name Shape Semantic Notes
query scalar —
patient_context map patient_context

Outputs

Name Shape Semantic Notes
answer map clinical_narrative

Tags: agent · cart

Runtime dependency

This agent is a Retrieval-Augmented Generation service: at runtime it needs a populated vector database and an LLM API key. When those are absent it returns an honest degraded response (e.g. HTTP 503) and never fabricates clinical content.

↩ Back to the Intelligence Agents index · the Capability Maturity Matrix · the Capability Brief.

Note

Status and interface are generated from the capability registry (lib/hcls_common/capabilities.json) — the site cannot claim ahead of the code. All clinical output is decision support for a qualified clinician, never autonomous diagnosis.