Spaces:
Sleeping
Sleeping
File size: 3,571 Bytes
774f1fa bfda151 e7279ce bfda151 e7279ce bfda151 e7279ce bfda151 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 | ---
title: Beacon Trial Finder
emoji: 📚
colorFrom: gray
colorTo: yellow
sdk: gradio
sdk_version: 6.14.0
python_version: '3.13'
app_file: app.py
pinned: false
license: mit
short_description: help patients with rare disease to find clinical trials
---
# Beacon — Rare Disease Clinical Trial Finder
Beacon is a conversational AI assistant that helps patients with rare diseases find relevant recruiting clinical trials near them. It conducts a warm intake interview, geocodes the patient's location, queries [ClinicalTrials.gov](https://clinicaltrials.gov) in real time, and produces a ranked report of the closest matching trials.
## How it works
1. **Intake agent** (Claude Sonnet) — interviews the patient conversationally to collect disease, age, symptom onset date, formal diagnosis date, location, preferred trial phases, and whether they are interested in Expanded Access Programs (EAP / compassionate use).
2. **Research agent** (Claude Opus) — searches ClinicalTrials.gov via the official v2 API for clinical trials and/or EAP listings, retrying with synonyms or wider radii if results are sparse, then outputs a ranked report with eligibility notes and next steps.
3. **LangGraph** orchestrates the two-node pipeline (intake → research).
### Clinical trial phases explained
| Phase | Focus | Typical size |
|---|---|---|
| **Early Phase 1** | First-in-human safety; tiny doses | ~10–15 people |
| **Phase 1** | Safe dosage range and side effects | 20–80 people |
| **Phase 2** | Does it work? Continued safety | 100–300 people |
| **Phase 3** | vs. standard of care; required for FDA approval | 1,000–3,000 people |
| **Phase 4** | Post-approval long-term surveillance | Varies |
### Expanded Access Programs (EAP)
EAP (also called compassionate use) allows patients who do not qualify for or cannot access a clinical trial to receive an investigational drug or device outside of a formal trial. The treatment is not yet FDA-approved; a physician must submit the EAP request to the drug sponsor and obtain FDA authorization. Beacon can search for available EAP listings alongside clinical trials.
## Project setup
### Prerequisites
- Python 3.11+
- [`uv`](https://docs.astral.sh/uv/) (recommended) or `pip`
### 1. Clone the repo
```bash
git clone <repo-url>
cd beacon
```
### 2. Install dependencies
```bash
uv sync
```
Or with pip:
```bash
pip install -e .
```
### 3. Configure API keys
Copy the example env file and fill in your keys:
```bash
cp .env.example .env
```
Open `.env` and replace the placeholder values:
```env
ANTHROPIC_API_KEY=your_anthropic_api_key_here
```
### 4. Run
```bash
uv run python app.py
```
#### Hot-reload during development
Use the `gradio` CLI to automatically reload the app whenever you save a file — no manual restart needed:
```bash
uv run gradio app.py
```
> **Note:** Active user sessions are reset on each reload.
## Configuration
| Environment variable | Default | Description |
|----------------------|-------------|-------------------|
| `ANTHROPIC_API_KEY` | *(required)*| Anthropic API key |
## Project structure
```
beacon/
├── app.py # Gradio web UI entry point
├── main.py # Terminal entry point
├── clinical_trials_guru.py # Intake + research agents, LangGraph pipeline
├── llm.py # LLM provider abstraction
├── requirements.txt # HuggingFace Spaces dependencies
├── pyproject.toml
└── .env # Local secrets — not committed
```
|