🔭 TokenLens Pro — All-in-One AI Usage, Quota & Cost Intelligence
"See the true cost, quota limits, and engineering impact of every AI model you use — in one place."
📸 Screenshots & Live Previews
1. Interactive AI Quotas & Cost Intelligence Dashboard
Track your real-time spend, session count, prompt cache hit rate %, and multi-window quota meters across all active AI coding tools:
2. Multi-Window Quota Tracking & Live Reset Countdowns (Antigravity, Codex, Cursor)
Monitor your exact remaining quota pools for Gemini Models (Weekly & 5-Hour limits), Claude/GPT models, and reset countdowns in real-time:
🌟 Why TokenLens?
Developers today use multiple AI tools concurrently — Google Antigravity, OpenAI Codex, Cursor, Claude Code, Windsurf, GitHub Copilot, Ollama, and DeepSeek. Managing their rate limits, prompt caching, unexpected credit exhaustion, and costs has been painful and fragmented.
TokenLens brings everything into one unified, ultra-fast, and private VS Code extension:
- ⏱️ Live Multi-Window Rate Limits & Quota Pools: Weekly and 5-hour limit meters with exact countdown timers (
2h 15m, 4d 12h).
- ⚡ Individual Provider Status Bar Indicators: Displays dedicated compact circular gauges for every active AI tool simultaneously (e.g.
● AG 45%, ● Codex 21%, ● Copilot ✓, ● Claude 18%).
- 🔍 In-Depth Hover Telemetry: Hover over any provider's status bar item to view model-by-model quota meters, 5-hour and weekly reset countdowns, and budget tracking.
- 🌐 Global Multi-Currency Conversion: Instantly convert usage and spend into USD ($), BDT (৳), EUR (€), INR (₹), GBP (£), JPY (¥), CAD, AUD.
- 💡 AI Cost Optimization Advisor: Actionable suggestions to downscale routine tasks to mini models and save up to 65%.
- 🚀 Prompt Cache Analyzer: Real-time analysis of cached vs uncached tokens, hit rate %, and dollar savings.
- 📈 Interactive Trend Analytics & Model Leaderboard: Interactive cost charts and usage split by model, provider, and project workspace.
- 💰 Smart Budget Guardrails: Custom monthly budget alerts at 75%, 90%, 95% (Panic Mode), and 100%.
- 🔒 100% Local & Zero Telemetry: Credential decoding and database queries run entirely on your local machine.
🤖 Supported Providers (11 Providers)
| Provider |
Mechanism & Strategies |
Quotas & Windows Monitored |
| Google Antigravity |
Language Server Port Scan (lsof), GetQuotaSummary, GetUserStatus, SQLite Protobuf decode (state.vscdb), Cloud Code API |
Gemini Models (Weekly & 5h limits) + Claude/GPT Models with live reset countdowns |
| OpenAI Codex |
Local session logs (~/.codex/sessions/), CLI auth token |
Session quota %, token breakdown, lifetime spend |
| Cursor |
SQLite DB (state.vscdb) token extraction, token refresh API, DashboardService/GetCurrentPeriodUsage |
Included usage %, Auto requests %, Fast requests, Plan spend $, Billing cycle reset |
| Windsurf |
SQLite state.vscdb, Dynamic listening port probe (lsof), LanguageServerService/GetUserStatus |
Available Prompt Credits, Used Prompt Credits, Flex Credits |
| Claude Code |
Credentials file (~/.claude/credentials.json), Session cookie, OAuth refresh |
5-Hour Session Window, 7-Day Weekly Window, Sonnet/Opus limits, Extra usage |
| GitHub Copilot |
VS Code GitHub Authentication & Copilot quota endpoints |
Premium request allowance, Chat suggestions, Monthly renewal date |
| DeepSeek |
API Key balance endpoint (api.deepseek.com/user/balance) |
Account balance in USD/CNY, Granted balance, Top-up balance |
| Mistral |
Admin Cookie / API key (console.mistral.ai/api/billing/usage) |
Usage credits, plan tiers |
| Ollama (Local) |
Local daemon API (localhost:11434/api/version, /api/tags, /api/ps) |
Server version, Installed model count, Running models, VRAM usage |
| OpenRouter |
API key endpoint (openrouter.ai/api/v1/auth/key, /api/v1/credits) |
Credit limit, Total USD usage, Rate limits (req/s) |
| Groq |
API key balance & rate limit headers |
Request quota, token rate limits |
⚡ Multi-Provider Sleek Status Bar
TokenLens gives each active provider its own compact circular indicator on your status bar:
● AG 45% ● Codex 21% ● Copilot ✓ ● Claude 18%
- Independent Multi-Tool Display: All active AI providers show side-by-side with crisp circular status indicators.
- Provider-Specific Hover Tooltip: Hovering over any item displays that exact provider's in-depth model limits, 5-hour and weekly reset timers, and budget status.
- Instant Click Action: Clicking any status bar item immediately opens and focuses the TokenLens dashboard.
⌨️ Keyboard Shortcuts & Commands
| Command |
Shortcut |
Description |
TokenLens: Refresh All Providers |
R (inside dashboard) |
Re-queries all language servers, APIs, and local logs |
TokenLens: Open Dashboard |
— |
Focuses and opens the TokenLens AI Dashboard |
TokenLens: Select Status Bar Provider |
— |
Pin a specific provider or set to Auto (all active) |
TokenLens: Set Monthly Budget |
— |
Configure your monthly spending budget in USD |
TokenLens: Set Hourly Rate for ROI |
— |
Set your developer hourly rate for ROI savings math |
TokenLens: Detect Local Sources |
— |
Auto-scans local directories for Claude, Codex, Grok, and Cline logs |
TokenLens: Export CSV / JSON |
— |
Export full audit logs and usage summaries |
⚙️ Configuration Options
Customize TokenLens in VS Code Settings (Cmd+, or Ctrl+,):
{
// Auto-refresh interval (e.g. 1m, 5m, 15m, 30m)
"tokenlens.refreshInterval": "5m",
// Monthly budget in USD for alert tracking
"tokenlens.budget.monthly": 20,
"tokenlens.budget.alertEnabled": true,
// Preferred display currency (USD, BDT, EUR, GBP, INR, JPY, CAD, AUD)
"tokenlens.display.currency": "USD",
// Status bar display style: "compact" | "blocks" | "percent" | "minimal"
"tokenlens.display.statusBarStyle": "compact",
// Pin a specific provider to the status bar (leave empty for auto multi-provider)
"tokenlens.display.statusBarProvider": "",
// Hourly rate for ROI calculation ($)
"tokenlens.roi.hourlyRate": 35,
// Ollama local server URL
"tokenlens.providers.ollama.url": "http://localhost:11434"
}
🔒 Privacy Guarantee
TokenLens is built from the ground up for security and privacy:
- ✅ 100% Local Processing: Runs on your machine; no code, prompts, or proprietary data ever leaves your device.
- ✅ Secure Secret Storage: API keys and credentials are saved using VS Code's encrypted SecretStorage.
- ✅ Zero Telemetry: No third-party trackers, analytics, or external logging servers.
🛠️ Building & Packaging from Source
git clone https://github.com/iamtashanto/TokenLens.git
cd TokenLens
pnpm install
pnpm run compile
pnpm run build
pnpm test
pnpm run package:vsix
📄 License
Distributed under the MIT License. Created with ❤️ for AI engineers and developers worldwide.
| |
