Ollama Request Interceptor & Log Viewer

This tool helps you intercept, log, and visualize requests and responses sent to a local Ollama server. It's useful for debugging, auditing, or understanding how prompts are being handled by your models.

🧰 Dependencies

• Go
• Python (3.7+)
• Streamlit

Install Python dependencies (once):

pip install streamlit

---

⚙️ Getting Started

1. Clone the Repository

2. Run the Interceptor

sh run-interceptor.sh

defaults: Target URL : http://localhost:11434 Log Dir : ./logs Listen Addr: :11435

or

sh run-interceptor.sh -target "http://ai-test-3:11434" -log "./newlogs" -listen ":7800"

• This starts a reverse proxy that logs requests/responses intended for the Ollama server.
• It creates a directory where logs are stored in a format the viewer understands.

---

📊 Viewing the Logs

Once you have logs:

3. Launch the Viewer

streamlit run olla.py

• Select the default logs/ folder or provide a custom path to view logs.
• Only logs written in the expected format (by the Go script) will be displayed correctly.

---

📌 Notes

• Make sure the client app is configured to point to the proxy (e.g., http://localhost:11435) instead of directly to Ollama.
• The proxy will forward the request and respond transparently, while logging everything in between.