Skip to content

Cerebral Courtroom

Build a real-time, multi-agent courtroom simulator powered by GPT-5-Codex that allows users to present legal arguments and receive dynamic feedback from AI agents playing the roles of Judge, Opposing Counsel, and Jury.

Code Video Cerebral Valley

Core trick

Runs judge, opposing-counsel, and jury agents in a FastAPI backend that responds to submitted legal arguments.

Tags

legalagents

Description

Build a real-time, multi-agent courtroom simulator powered by GPT-5-Codex that allows users to present legal arguments and receive dynamic feedback from AI agents playing the roles of Judge, Opposing Counsel, and Jury. The system predicts trial outcomes and highlights persuasive strategies.
panforrest/cerebral-courtroom★ 00 forksPythonCreated Sep 28, 2025 · last push Sep 28, 2025

README excerpt

# Cerebral Courtroom

MVP scaffold for Cerebral Courtroom — a multi-agent courtroom simulation using GPT-5-Codex.

Quickstart

1. Create a Python virtual environment and activate it.

2. Install backend dependencies:

   ```powershell
   python -m venv .venv
   .\.venv\Scripts\activate
   pip install -r backend/requirements.txt
   ```

3. Set your OpenAI API key in the environment:

   ```powershell
   $env:OPENAI_API_KEY = "sk-..."
   ```

4. Run the backend:

   ```powershell
   uvicorn backend.main:app --reload --port 8000
   ```

   Demo endpoint (single-agent opposing counsel):

   1. Start the backend as above.
   2. POST to `http://localhost:8000/api/demo/opposing` with JSON:

   ```powershell
   curl -X POST "http://localhost:8000/api/demo/opposing" -H "Content-Type: application/json" -d '{"facts": "Alice saw Bob at the store.", "argument": "The defendant was at the scene."}'
   ```

   This returns a JSON `reply` from the Opposing Counsel agent.

Frontend and further instructions will be added during the hackathon.

Running end-to-end Playwright tests

- Mocked (default, safe for CI):

   ```powershell
   # from the project root
   python -m pytest tests/e2e/test_demo_playwright.py -q -s
   ```

- Real OpenAI API (opt-in):

   ```powershell
   # set your API key and enable real API mode
   $env:OPENAI_API_KEY = "sk-..."
   $env:DEMO_USE_REAL_API = "1"
   python -m pytest tests/e2e/test_demo_playwright.py -q -s
   ```

   Or use pytest CLI flag (sets the env var for y

Team