CAPE Sandbox Orchestrator API

This project exposes a small FastAPI service that submits files or URLs to a CAPE sandbox, polls the analysis task until the report is ready, stores the result in SQLite, and optionally asks a local Ollama endpoint for a short verdict.

What it does

• Accepts a file upload or URL through POST /submit
• Submits the sample to CAPE
• Polls the CAPE task until the report is generated
• Fetches the final CAPE report and screenshots
• Stores the analysis result in analysis_results.db
• Returns stored results through GET /results/{identifier}
• Sends a compact report summary to Ollama for a JSON verdict

CAPE documentation

Official CAPE documentation: https://capev2.readthedocs.io/

Setup

1. Create and edit the .env file if needed:

CAPE_URL=http://localhost:8000
DB_NAME=analysis_results.db
OLLAMA_URL=http://localhost:11434/api/generate

2. Install dependencies:

pip install -r requriments.txt

3. Run the API:

uvicorn CAPE_analysis:app --reload

Notes

• The code expects a running CAPE instance and, if you want AI verdicts, a running Ollama server.
• The SQLite database and other local runtime files are ignored by git so the repository stays clean.