claude-view
10 Claude sessions running. What are they doing? Live dashboard — monitor, cost tracking, search, sub-agent visibility.
claude-view
You have 10 Claude sessions running right now. What are they doing?
English · 繁體中文 · 简体中文 · 日本語 · Español · Français · Deutsch · Português · Italiano · 한국어 · Nederlands
Behind every "thinking..." spinner, Claude is spawning sub-agents, calling MCP servers, running skills, firing hooks — and you can't see any of it. You Cmd+Tab through 15 terminals trying to remember which session was doing what. A cache expired while you weren't looking. A session finished 10 minutes ago and you didn't notice. Another one hit its context limit and you're burning tokens on a dead conversation.
You're paying $200/mo for Claude Code. You deserve a dashboard.
curl -fsSL https://get.claudeview.ai/install.sh | sh
One command. Every session visible. Real-time.
What You Get
Live Monitor
| Feature | Why it matters |
|---|---|
| Session cards with last message | Instantly remember what each long-running session is working on |
| Notification sounds | Get pinged when a session finishes or needs your input — stop polling terminals |
| Context gauge | Real-time context window usage per session — see which ones are in the danger zone |
| Cache warm countdown | Know exactly when prompt cache expires so you can time your next message to save tokens |
| Cost tracking | Per-session and aggregate spend — hover for token/cost breakdown with cache savings by category |
| Sub-agent visualization | See the full agent tree — sub-agents, their status, and what tools they're calling |
| Multiple views | Grid, List, Kanban, or Monitor mode — pick what fits your workflow |
| Kanban swimlanes | Group sessions by project or branch — visual swimlane layout for multi-project workflows |
| Recently closed sessions | Sessions that end appear in a "Recently Closed" section instead of vanishing — persist across server restarts |
| Unified live chat | History and real-time messages in a single scrollable conversation — no more tab-switching |
| Queued message indicators | Messages waiting in the queue show as pending bubbles with a "Queued" badge |
| SSE-driven live data | Live session data pushed via Server-Sent Events — eliminates stale-cache risks entirely |
Rich Chat History
| Feature | Why it matters |
|---|---|
| Full conversation browser | Every session, every message, fully rendered with markdown and code blocks |
| Tool call visualization | See file reads, edits, bash commands, MCP calls, skill invocations — not just text |
| Compact / verbose toggle | Skim the conversation or drill into every tool call |
| Thread view | Follow agent conversations with sub-agent hierarchies |
| Export | Markdown export for context resumption or sharing |
| Bulk select & archive | Select multiple sessions for batch archiving with persistent filter state |
| Encrypted sharing | Share any session via E2E encrypted link — zero server trust |
Advanced Search
| Feature | Why it matters |
|---|---|
| Full-text search | Search across all sessions — messages, tool calls, file paths |
| Unified search engine | Tantivy full-text + SQLite pre-filter run in parallel — one endpoint, no frontend wiring gaps |
| Project & branch filters | Scope to the project you're working on right now |
| Command palette | Cmd+K to jump between sessions, switch views, find anything |
Plans, Prompts & Teams
| Feature | Why it matters |
|---|---|
| Plan browser | View your .claude/plans/ directly in session detail — no more hunting through files |
| Prompt history | Full-text search across all prompts you've sent to Claude Code with template clustering and intent classification |
| Teams dashboard | See team leads, inbox messages, team tasks, and file changes across all team members — with JSONL fallback so history is never lost |
| Prompt analytics | Leaderboards of prompt templates, intent distribution, and usage statistics |
Plugin Manager
| Feature | Why it matters |
|---|---|
| GUI plugin browser | Browse, install, enable, disable, and uninstall Claude Code plugins without touching a terminal |
| Plugin card grid | Every installed plugin shown as a card with name, status, scope, and version |
| Marketplace dialog | Discover new plugins and install with scope picker (user-level or project-level) |
| Health banner | Plugin system status at a glance — know when something is broken before it breaks your workflow |
Workflows
| Feature | Why it matters |
|---|---|
| Workflow builder | Create and run multi-stage workflows from a dedicated page — VS Code-style layout with Mermaid diagram preview and YAML editor |
| Streaming LLM chat rail | Generate workflow definitions in real time using an embedded chat interface |
| Stage runner | Visualize stage columns, attempt cards, and progress bar as your workflow executes |
| Built-in seed workflows | Plan Polisher and Plan Executor ship out of the box |
Open in IDE
| Feature | Why it matters |
|---|---|
| One-click file open | Files referenced in sessions have a button to open them directly in your editor |
| Auto-detects your editor | VS Code, Cursor, Zed, and others auto-detected — no configuration needed |
| Everywhere it matters | Button appears in Changes tab, file headers, and Kanban project headers |
| Preference memory | Your preferred editor is remembered across sessions |
Agent Internals — See What's Hidden
Claude Code does a lot behind "thinking..." that never shows in your terminal. claude-view exposes all of it.
| Feature | Why it matters |
|---|---|
| Sub-agent conversations | See the full tree of spawned agents, their prompts, and their outputs |
| MCP server calls | See which MCP tools are being invoked and their results |
| Skill / hook / plugin tracking | Know which skills fired, which hooks ran, what plugins are active |
| Hook event recording | Dual-channel hook capture (live + JSONL backfill) — every hook event recorded and browsable, even for past sessions |
| Worktree branch drift | Detects when git worktree branches diverge — shown in live monitor and history |
| @File mention chips | @filename references extracted from sessions shown as chips on cards — hover for the full path |
| Agent SDK live chat | Thinking blocks, tool calls, and tool results with syntax highlighting, collapsible sections, and a context window gauge |
| Tool use timeline | Action log of every tool_use/tool_result pair with timing |
| Error surfacing | Errors bubble up to the session card — no more buried failures |
| Raw message inspector | Drill into any message's raw JSON when you need the full picture |
Analytics
A rich analytics suite for your Claude Code usage. Think Cursor's dashboard, but deeper.
Dashboard Overview| Feature | Description |
|---|---|
| Week-over-week metrics | Session count, token usage, cost — compared to your previous period |
| Activity heatmap | 90-day GitHub-style grid showing your daily Claude Code usage intensity |
| Top skills / commands / MCP tools / agents | Leaderboards of your most-used invocables — click any to search matching sessions |
| Most active projects | Bar chart of projects ranked by session count |
| Tool usage breakdown | Total edits, reads, and bash commands across all sessions |
| Longest sessions | Quick access to your marathon sessions with duration |
| Feature | Description |
|---|---|
| Code output tracking | Lines added/removed, files touched, commit count — across all sessions |
| Cost ROI metrics | Cost per commit, cost per session, cost per line of AI output — with trend charts |
| Model comparison | Side-by-side breakdown of output and efficiency by model (Opus, Sonnet, Haiku) |
| Learning curve | Re-edit rate over time — see yourself getting better at prompting |
| Branch breakdown | Collapsible per-branch view with session drill-down |
| Skill effectiveness | Which skills actually improve your output vs which ones don't |
| Feature | Description |
|---|---|
| Pattern detection | Behavioral patterns discovered from your session history |
| Then vs Now benchmarks | Compare your first month to recent usage |
| Category breakdown | Treemap of what you use Claude for — refactoring, features, debugging, etc. |
| AI Fluency Score | Single 0-100 number tracking your overall effectiveness |
Insights and Fluency Score are in early experimental stage. Treat as directional, not definitive.
Built for the Power User
claude-view is for the developer who:
- Runs 3+ projects simultaneously, each with multiple worktrees
- Has 10-20 Claude Code sessions open at any time
- Needs to context-switch fast without losing track of what's running
- Wants to optimize token spend by timing messages around cache windows
- Gets frustrated by Cmd+Tab-bing through terminals to check on agents
- Worktree-aware — detects branch drift across git worktrees
How It's Built
| Blazing fast | Rust backend with SIMD-accelerated JSONL parsing, memory-mapped I/O — indexes thousands of sessions in seconds |
| Real-time | File watcher + SSE + unified WebSocket with heartbeat, event replay, and crash recovery |
| Tiny footprint | ~10 MB download, ~27 MB on disk. No runtime dependencies, no background daemons |
| 100% local | All data stays on your machine. Zero telemetry, zero required accounts. Optional encrypted sharing available. |
| Zero config | npx claude-view and you're done. No API keys, no setup, no accounts |
Measured on an M-series Mac with 1,493 sessions across 26 projects:
| Metric | claude-view | Typical Electron dashboard |
|---|---|---|
| Download | ~10 MB | 150–300 MB |
| On disk | ~27 MB | 300–500 MB |
| Startup (ready to serve) | < 500 ms | 3–8 s |
| RAM (full index loaded) | ~50 MB | 300–800 MB |
| Deep-index 1,500 sessions | < 1 s | N/A |
| Runtime dependencies | 0 | Node.js + Chromium |
Key techniques:
- SIMD pre-filter —
memchrscans raw bytes before touching a JSON parser - Memory-mapped I/O — JSONL files are mmap'd and parsed in-place, never copied
- Tantivy search — Same engine behind Quickwit; indexes 1,500 sessions in under a second
- Zero-copy where it counts — Borrowed slices from mmap through parse to response
cargo build --release
/usr/bin/time -l target/release/claude-view # peak RSS + wall time
How It Compares
The Claude Code ecosystem has great tools — chat UIs, history viewers, session managers. claude-view fills a different gap: real-time monitoring + deep history + analytics in one lightweight workspace.
| Tool | Category | Stack | Size | Live monitor | Search | Analytics |
|---|---|---|---|---|---|---|
| claude-view | Monitor + workspace | Rust | ~10 MB | Yes | Yes | Yes |
| opcode | GUI + session manager | Tauri 2 | ~13 MB | Partial | No | Yes |
| ccusage | CLI usage tracker | TypeScript | ~600 KB | No | No | CLI |
| CodePilot | Desktop chat UI | Electron | ~140 MB | No | No | No |
| claude-run | History viewer | TypeScript | ~500 KB | Partial | Basic | No |
Why the size difference mattersChat UIs (CodePilot, CUI, claude-code-webui) are interfaces for Claude Code. claude-view is a dashboard that watches your existing terminal sessions. They're complementary, not competing.
| claude-view | Electron app | |
|---|---|---|
| Download | ~10 MB | ~140 MB |
| On disk | ~27 MB | ~400 MB |
| What's in it | Rust server + SPA assets | Chromium + Node.js + Next.js + app code |
| RAM at idle | ~50 MB | ~300 MB+ |
| Startup | < 500 ms | 3–8 s |
| Background cost | Negligible | Chromium renderer process |
When you're already running 10+ Claude Code sessions eating RAM and CPU, the last thing you want is a 300 MB dashboard competing for resources.
Installation
| Method | Command |
|---|---|
| Shell (recommended) | curl -fsSL https://get.claudeview.ai/install.sh | sh |
| npx | npx claude-view |
The shell installer downloads a pre-built binary (~10 MB), installs to ~/.claude-view/bin, and adds it to your PATH. Then just run claude-view.
Only requirement: Claude Code installed — this creates the session files we monitor.
Configuration| Env Variable | Default | Description |
|---|---|---|
CLAUDE_VIEW_PORT or PORT |
47892 |
Override the default port |
The pre-built binary (npx claude-view) ships with auth, sharing, and mobile relay baked in. If you're building from source or running your own server, these features are opt-in via environment variables — omit any of them and that feature is simply disabled. No errors, no warnings, just graceful degradation.
| Env Variable | Feature | What happens without it |
|---|---|---|
SUPABASE_URL |
Login / auth | Auth disabled — app runs in fully local, zero-account mode |
RELAY_URL |
Mobile pairing (QR code) | Mobile relay disabled — QR pairing unavailable |
SHARE_WORKER_URL + SHARE_VIEWER_URL |
Encrypted session sharing | Share button hidden — sessions stay local |
To run fully local with no cloud dependencies (the default for self-hosters):
bun dev # just works — all cloud features off
To enable a feature, export the env var before starting the server:
export SUPABASE_URL=https://your-project.supabase.co
export RELAY_URL=wss://your-relay.fly.dev/ws
export SHARE_WORKER_URL=https://your-worker.example.com
export SHARE_VIEWER_URL=https://your-viewer.example.com
bun run dev:server
Corporate / Sandbox EnvironmentsThese are runtime overrides — they take precedence over whatever is baked into the binary at build time.
If your machine restricts writes to ~/Library/Caches/ (e.g., DataCloak, CrowdStrike, corporate DLP):
cp crates/server/.env.example .env
# Uncomment the CLAUDE_VIEW_DATA_DIR line
This keeps the database, search index, and lock files in .data/ inside the repo — no writes outside the project directory.
FAQ
"Not signed in" banner showing even though I'm logged inclaude-view checks your Claude credentials by reading ~/.claude/.credentials.json (with macOS Keychain fallback). If it can't find or parse valid credentials, the warning banner appears. Try these steps in order:
1. Verify Claude CLI auth works
claude auth status
If this says "Logged in", your credentials are valid. If not, run claude auth login first.
2. Check the credentials file exists
cat ~/.claude/.credentials.json
You should see JSON with a claudeAiOauth section containing an accessToken. If the file is missing, your credentials may be stored only in the OS Keychain (see step 3).
3. Check macOS Keychain access
security find-generic-password -s "Claude Code-credentials" -w
If this returns JSON or hex-encoded data, credentials are in Keychain. If it returns an error, or you're on a corporate machine with security tools (DataCloak, CrowdStrike, etc.) that block Keychain access, this is likely the issue — run claude auth login to regenerate the file.
4. Check for token expiry
Inside the credentials JSON, look at the expiresAt field (Unix milliseconds). If it's in the past, your token has expired. Run claude auth login to refresh.
5. Check HOME environment
claude-view reads credentials from $HOME/.claude/.credentials.json. If the server process has a different $HOME than your shell, it won't find your credentials.
echo $HOME
If all the above checks pass and the banner still shows, please report it on our Discord with the output of claude auth status and we'll help debug it.
Community
- Website: claudeview.ai — docs, changelog, blog
- Discord: Join the server for support, feature requests, and discussion
Development
Prerequisites: Rust, Bun, cargo install cargo-watch
bun install # Install all workspace dependencies
bun dev # Start full-stack dev (Rust + Web with hot reload)
Workspace Layout
| Path | Package | Purpose |
|---|---|---|
apps/web/ |
@claude-view/web |
React SPA (Vite) — main web frontend |
apps/share/ |
@claude-view/share |
Share viewer SPA (Vite) — Cloudflare Pages |
apps/mobile/ |
@claude-view/mobile |
Expo native app |
apps/landing/ |
@claude-view/landing |
Static HTML landing page |
packages/shared/ |
@claude-view/shared |
Shared types & theme tokens |
packages/design-tokens/ |
@claude-view/design-tokens |
Colors, spacing, typography |
packages/plugin/ |
@claude-view/plugin |
Claude Code plugin (MCP server + tools + skills) |
crates/ |
— | Rust backend (Axum) |
infra/share-worker/ |
— | Cloudflare Worker — share API (R2 + D1) |
infra/install-worker/ |
— | Cloudflare Worker — install script proxy with download tracking |
Dev Commands
| Command | Description |
|---|---|
bun dev |
Full-stack dev — Rust server (cargo-watch) + Web frontend (Vite HMR) |
bun run dev:web |
Web frontend only (assumes Rust server already running) |
bun run dev:server |
Rust backend only (with cargo-watch) |
bun run dev:all |
All JS/TS apps via Turbo (web + mobile + landing, no Rust) |
bun run build |
Build all workspaces |
bun run preview |
Build web + serve via release binary |
bun run lint |
Lint all JS/TS workspaces |
bun run lint:all |
Lint JS/TS + Rust (Clippy) |
bun run typecheck |
TypeScript type checking |
bun run test |
Run all tests (Turbo) |
bun run test:rust |
Run Rust tests (cargo test --workspace) |
cd apps/web && bunx vitest run |
Run web frontend tests only |
bun run fmt |
Format Rust code |
bun run cleanupport |
Kill processes on dev ports (47892, 5173) |
Testing Production Distribution:
bun run dist:test # One command: build → pack → install → run
| Command | Description |
|---|---|
bun run dist:pack |
Package binary + frontend into tarball at /tmp/ |
bun run dist:install |
Extract tarball to ~/.cache/claude-view/ (simulates first-run download) |
bun run dist:run |
Run the npx wrapper using the cached binary |
bun run dist:clean |
Remove all dist cache and temp files |
Releasing:
bun run release # patch bump: 0.1.0 → 0.1.1
bun run release:minor # minor bump: 0.1.0 → 0.2.0
bun run release:major # major bump: 0.1.0 → 1.0.0
git push origin main --tags # triggers CI → builds all platforms → auto-publishes to npm
Platform Support
| Platform | Status |
|---|---|
| macOS (Apple Silicon) | Available |
| macOS (Intel) | Available |
| Linux (x64) | Planned |
| Windows (x64) | Planned |
Related
- claudeview.ai — Official website, docs, and changelog
- @claude-view/plugin — Claude Code plugin with 8 MCP tools and 3 skills.
claude plugin add @claude-view/plugin - claude-backup — Claude Code deletes your sessions after 30 days. This saves them.
npx claude-backup
Yorumlar (0)
Yorum birakmak icin giris yap.
Yorum birakSonuc bulunamadi