Live Thought & Session Inspector
Send a prompt from Cursor or select a past session from the sidebar to stream internal reasoning traces and function calls.
Endpoint: /v1/chat/completions · Model: Gemini / GPT / Codex
--
--
--
--
•
--
Latency
0 ms
Speed
0.0 t/s
Tokens (In / Out)
0 / 0
Model Thought Trace (Internal Reasoning)
Raw model internal thinking stream captured directly before Cursor filtering
0 tokens
No thinking trace recorded for this request.
Assistant Response Content
Generated visible content returned to client
Waiting for assistant output...
Tool Calls & Function Invocations
Generated function calls dispatched to tools
Input Conversation Context
Full prompt history sent by Cursor/client (0 messages)
Click to expand