Read .jsonl transcripts like a conversation.

LLM session logs are unreadable raw — one JSON blob per line. Drop one here and read it as a chat: tool calls folded, thinking blocks tucked away, subagent chatter marked. Parsing happens in your browser; transcripts are sensitive and this page has no upload endpoint to leak them to.

understands: Claude Code session logs · OpenAI chat / fine-tune files · {prompt, reply} eval/bench logs · generic {role, content} lines