ChatTCP AI Diagnostic Analysis Assistant is Now Live

ChatTCP online version now features an AI diagnostic analysis assistant. We can have conversations with the AI assistant to help us analyze problems and data - through chatting, we can find solutions to issues.

ChatTCP parses data packets into AI-readable and understandable JSON structured data, providing annotations to help AI understand each packet, decode application layer protocols, and attach ChatTCP's preliminary diagnostic results and analysis data. Therefore, AI can understand network packet files.

The AI diagnostic analysis assistant doesn't use MCP, but rather Tool & Function Calling (method invocation), though the core logic still follows MCP principles. AI functionality is more convenient than using MCP because you can directly invoke the AI assistant to ask questions while using ChatTCP to view data packets, without having to switch to other AI tools.

ChatTCP's AI assistant currently integrates with OpenRouter service provider, using the Claude 3.7 Sonnet model. We will further seek models with better analysis performance in the future. Since OpenRouter doesn't support custom knowledge bases, we're also working on solutions or looking for other supporting service providers.

The AI diagnostic analysis assistant supports multi-turn Q&A conversations. The quality of AI responses is provided by Claude 3.7 Sonnet, and users need to judge for themselves whether the answers are accurate and correct.

For early subscribers, we have allocated 60 user message credits for annual subscribers and 300 user message credits for permanent subscribers to chat with the AI assistant. This is determined based on early subscription pricing.

Credits can be recharged after use. Currently, the recharge function is not yet launched, so recharge is temporarily unavailable, and recharge pricing is undetermined.