Approval Radar
Approval Radar is an Astra-powered research agent that helps biotech investors understand what could support or derail a new medicine’s FDA approval.
Core trick
Agent chooses regulatory and clinical-trial database queries, reads primary documents, and tests whether evidence supports approval, producing a cited brief.
Description
Approval Radar is an Astra-powered research agent that helps biotech investors understand what could support or derail a new medicine’s FDA approval.
Astra chooses research tools, queries regulatory and clinical-trial databases, reads original documents, and tests whether the evidence supports the approval case. It connects to FDAgent’s existing data tools and produces a cited brief covering clinical results, safety, manufacturing and regulatory risks.
In one case, Astra found that a reassuring inspection covered research practices and did not resolve the drug’s manufacturing concern. It identified the exact evidence still needed.
We completed 35 real Astra runs across 33 drug cases. Users can inspect the research trail, challenge conclusions and export findings.
README excerpt
# FDAgent Approval Radar **Investigate what could change an FDA drug-review outcome, using public evidence and GPT-6 Astra.** Approval Radar follows **33 real drug–indication review episodes across 29 companies**, with **98 curated primary sources** and recorded public-database scans for every candidate. It connects sponsor disclosures, trial protocols and results, FDA reviews and letters, approved applications, labeling and research publications. Astra investigates which evidence changes the case and turns that assessment into specific diligence questions. No customer uploads are required. All **33 candidates have source-reviewed deep Astra investigations**, with clinical, safety, manufacturing and regulatory assessments. The 35 recorded runs include repeat challenges; they are not 35 different drugs. [Coverage and individual reports](docs/research-coverage.md) distinguish supportive evidence, concerns and unresolved questions. Two reports carry explicit editorial source-review qualifications in the app and exported briefs; the original model outputs remain unchanged. Built for the GPT-6 Astra hackathon in New York. This is an original, standalone public project; the private FDAgent application and its data are not included.  ## Try it - [Public interactive demo](https://kllymx.github.io/fdagent-approval-radar/) — genuine recorded Astr