CS
ClaudeSense
100% local, 100% offline · Reads only ~/.claude/projects/ — no network calls, no telemetry
100% Free · Forever

Make sense of every Claude Code session

Token cost, habits, coaching, and visibility — all from your local session logs. A free VS Code extension that reads ~/.claude/projects/ and never sends a byte anywhere else.

Editorial illustration of a ClaudeSense dashboard window — a list of sessions with A/B/C health grades, a health-score ring at 85%, and cache hit, corrections, and context metrics — surrounded by labelled chips for Token Cost, Coaching, Visibility, and Habits

A single pane for every session on your disk.

ClaudeSense sidebar showing 58 sessions across 21 projects with total estimated cost, a 4-way date filter (Today, 7d, 30d, All), and each session row showing its health grade and token count
Every session, one list. Grades, tokens, and cost at a glance.
Health factor bars for a session — cache efficiency, accuracy, and code awareness — with tiered coaching advice expanded underneath, explaining what each percentage range means and what to do about it
Coaching that explains itself. Every factor comes with plain-language advice, not just a score.
Context window hero card showing 68.4% full against a 1M-token limit, with safe ceiling, turns left, and average per-turn stats, plus a chronological turn table tagged with VERBOSE flags and cost pills
See where your tokens go. Context window, top turns, and flags like VERBOSE and CORRECTION — turn by turn.

Tips you can act on — with the numbers to back them up.

Tips panel showing two suggestions — verbose responses and blind edits detected — followed by token and tool breakdown panels with cache hit rate at 88.4%, and a context window bar at 42.4% full

Not just a score. Nudges point to the exact turns, files, and patterns that pulled your health down — so you know what to change next session.

Four things you'll see, every session.

Token Cost

Every turn tallied — input, cache, output. Dollar estimates use public API rates, so they're directional, not a bill. Enough signal to see which habits actually burn tokens.

Tokens first. Dollars second.

Coaching

Inline tips based on the patterns Claude Code already shows. Correction rate climbing? Editing files without reading first? You'll see the advice before the bad habit sticks.

Every session gets a tip. Read it, or don't.

Visibility

Cache hit rate, context window pressure, tools used, and a 90-day activity heatmap. Everything the JSONL knows, surfaced without you asking.

Every token, every tool call — logged locally.

Habits

VERBOSE, UNCACHED, CORRECTION, SUB-AGENT. Flags mark the turns worth noticing so you can see what a wasteful session actually looks like.

Learn what your wasteful turns look like.
~5.1M
tokens
≈ $3.24 est.

Measured where the action is.

Tokens first, dollars second. Cost is an estimate, never a bill.

100% local, 100% offline, 100% free.

ClaudeSense reads only your local ~/.claude/projects/ files — the same session logs Claude Code already writes to your disk. No network calls. No telemetry. No tokens consumed by the extension itself.

Offline · No API calls

Questions worth answering.

ClaudeSense is free. The only costs are the ones Anthropic already bills you for using Claude Code itself — the extension doesn't add any.
Yes — it reads sessions regardless of how you're billed. But the dollar estimates use public API rates, so they won't match your actual Pro/Max subscription bill. Treat the cost numbers as directional.
No. Zero network calls in the extension. It reads local files, renders a dashboard, writes a pricing config to your home directory. That's the whole I/O surface.
No — Claude Code only. It parses the specific JSONL format Claude Code writes; other tools use different logs (or none at all).
Monthly cadence. Updates arrive through the VS Code Marketplace automatically — no action needed. The full release history is on the Marketplace listing.
Yes. Edit ~/.claude/claudesense-pricing.json — rates are per million tokens, and the file regenerates with sensible defaults if you delete it.
IS
Ishdeep S Sahni
Senior Staff PM · AI Workflow & Skills Coach

Ishdeep is a Senior Staff Product Manager and UX Lead with 18+ years in design and product. He works at the intersection of enterprise product design and AI integration — building AI-assisted design workflows with Claude and Figma MCPs, and coaching product managers and designers on using AI well. ClaudeSense is an extension of that work: making AI-assisted development observable.

Install free from the Marketplace.

Updates arrive automatically — every month.

Install from Marketplace