Tokentap is an open-source GitHub project whose title says it can print what an LLM is “thinking.” It was submitted to Hacker News, where the item currently has a score of 2 and one comment. The available listing does not specify which models or APIs are supported, how output is captured, whether the output represents hidden reasoning or merely streamed tokens, or what safety boundaries exist. It is therefore best treated as a project to inspect and validate, rather than as established evidence of reliable reasoning observability.
No heat snapshots are available in the last 24 hours.