SYS.TOOLS
Drop a Claude Code session JSONL file and explore its messages, tool calls, token usage, and estimated cost. 100% client-side: nothing leaves your browser.
Last updated
Drop session .jsonl files here, or click to browse
Add the subagents/ files alongside the session file to nest subagents under the call that spawned them.
100% client-side. Files never leave your browser.
Claude Code session files live at ~/.claude/projects/<project>/<session-id>.jsonl
Claude Code saves every session as a JSONL file under ~/.claude/projects/, in a folder named after your project's path. Drop any of those .jsonl files here to browse the prompts, tool calls, and results. Subagents live in a separate subagents/ folder beside the session file: select both and they nest under the call that spawned them.
Yes. This tool reads the usage data recorded in your session file and shows total input, output, and cache tokens, plus a per-turn breakdown and an estimated API cost. It correctly deduplicates streamed response chunks so token counts are accurate.
No. Parsing happens entirely in your browser with no API call and no server upload. Session files contain your code, file paths, and sometimes secrets, which is exactly why this tool is 100% client-side.
Token counts from the session file are multiplied by Anthropic's public API list prices for the models used, including separate rates for cache reads and cache writes. It is an estimate at API prices: if you use a Pro or Max subscription, you are not billed per token.
The whole session as a navigable tree: every tool call with its input, output, and duration, thinking blocks, per-turn token usage, and timing bars that make slow tool calls obvious at a glance. Drop the subagents/ files in with the session file and each subagent transcript nests under the call that spawned it.
