claudex

Appearance

claudexQuery your Claude Code sessions

A Rust CLI that indexes every JSONL transcript Claude Code writes under ~/.claude/projects/ and turns them into reports — cost, tools, turns, PRs, full-text search, and more.

Get Started →
Commands
GitHub

Every session, indexed

Scans ~/.claude/projects/ into a local SQLite database at ~/.claudex/index.db. Incremental sync keyed on (path, size, mtime); staleness window is 5 minutes.

Ten reports out of the box

summary, sessions, cost, tools, models, turns, files, prs, search, export — plus a live log watcher. Every read command supports --json and --no-index.

Honest pricing math

Separate Opus / Sonnet / Haiku tiers applied per-message from the model field in each record. Sub-cent values fall back to four decimals so tiny sessions don't round to $0.00.

Worktree-aware

Sessions inside .claude/worktrees/<branch> aggregate to the parent project and render as "project (worktree)". Group-by-project queries do the right thing automatically.

FTS5 full-text search

The index ships with a messages_fts virtual table. Search across every user and assistant message in every session, filtered by project, with SQLite's FTS5 ranking.

Live tail with structure

claudex watch tails Claude Code's --debug-file log in real time, formats tool calls, detects new sessions, and separates them with a banner. --raw drops back to plain output.

JSON or TTY — your choice

Human output uses a minimal comfy-table layout with dynamic width detection. --json emits a stable shape for pipelines, grep, and jq.

Single binary, no daemon

Built with rusqlite (bundled), clap, and owo-colors. Runs on Linux and macOS. Nix flake included. cargo install or build straight from source.

At a glance ​

$ claudex summary

# Top 5 projects by cost, last 30 days
$ claudex cost --limit 5

# Which files get touched most across all my projects?
$ claudex files --limit 10

# Find every session where I discussed "migrations"
$ claudex search migrations

# Export one session as Markdown
$ claudex export <session-prefix> --output session.md

Install ​

bash
# With Cargo
cargo install --git https://github.com/utensils/claudex

# Or with Nix flakes
nix run github:utensils/claudex -- summary

# Or from a local checkout
git clone https://github.com/utensils/claudex
cd claudex && nix develop && cargo build --release

See the full installation guide for Nix, devshell, and shell-completion setup.

Why claudex? ​

Claude Code persists every conversation as JSONL under ~/.claude/projects/<encoded-path>/<session>.jsonl. That's a gold mine — it records every user turn, every assistant message, every tool call, every token-usage block, every file modification — but those files are flat logs, not a queryable store.

claudex reads them once, indexes the parts you actually want to ask questions about, and gives you a CLI that answers questions like:

  • Which project burned the most Opus tokens last week?
  • What's my p95 turn duration in this repo?
  • Show me every session that linked a PR.
  • Full-text search: where did I first discuss the schema migration?
  • How many times have I edited src/index.rs across sessions?

No cloud. No daemon. No background service. Just a small Rust binary and a SQLite file under ~/.claudex/.

Next steps ​

Released under the MIT License.

Copyright © James Brink