A developer has outlined a method for programmatically capturing LLM chat responses across various providers, distinguishing between direct API calls and the responses shown to end-users in chat applications. For API interactions, tools like LiteLLM and OpenTelemetry are recommended for logging inputs, outputs, and exceptions. To capture user-facing responses from applications like ChatGPT and Gemini, a browser automation approach is necessary, as these platforms do not expose their specific model names or internal logic through standard APIs. The proposed solution involves using a normalized record schema to store details such as the provider, prompt, country, timestamp, model, text, markdown, and sources for each captured response. AI
IMPACT Provides a technical solution for developers needing to log and compare LLM outputs across different platforms.
RANK_REASON The item describes a technical method and tools for capturing LLM responses, not a new model release or significant industry event.
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →