Troubleshooting

This page helps resolve common BurpGPT Pro setup and provider issues.

Start with the simplest check that matches the symptom. The extension logs provider requests and responses in Burp's Logger tab, which is usually the fastest way to identify configuration problems.

The extension loads, but Pro features are not visible

Check that:

• You are using Burp Suite Professional 2026.4 or later.

• Your licence is active.

• Burp Suite was restarted after loading or updating the extension.

If activation fails, see What are the possible causes for an unsuccessful activation of my BurpGPT Pro?.

For the activation steps, see Activate your licence.

Test request fails

Check the provider response in Burp's Logger tab, then verify:

• The selected provider matches the endpoint you configured.

• The Base URL is correct.

• The API key is present when the provider requires one.

• The Model value is an exact model identifier supported by the provider.

• Request parameters is valid JSON and only contains parameters supported by that provider.

• Your provider account has enough credits, quota, or local resources to answer the request.

For the full validation flow, see Test and validate model provider settings.

Local provider requests do not connect

For Ollama, confirm that Ollama is running and that the model appears in:

ollama list

If Ollama is running on another host or port, update Base URL in Provider settings.

For OpenAI-compatible APIs, confirm that the configured Base URL usually ends in /v1 and that the endpoint supports chat completion requests.

Scans do not produce results

Check that:

• Test request succeeds with the active provider.

• The AI scanner prompt includes the placeholders you expect, such as {REQUEST} and {RESPONSE}.

• The Scan: On toggle is enabled before sending traffic to Burp's active scanner.

• For targeted analysis, you used Extensions -> BurpGPT Pro -> Scan with AI (results in Target -> Site map).

Scan-based results appear as Information-level issues named GPT-generated insights.

AI Chat does not include the expected traffic

In the AI Chat editor, keep Request, Response, or both selected before sending the message.

After selected traffic is sent once, the extension clears the toggle so follow-up messages do not repeatedly include the same data.