Tokentap

Free

Intercept LLM API traffic and visualize token usage in a real-time terminal dashboard. Track costs, debug prompts, and monitor context window usage across your AI development sessions.

FreeFree tier

#Python

Type

Saas

Visit Website GitHub (794 stars)

About Tokentap

Tokentap (formerly Sherlock) is a token tracker for LLM CLI tools that provides a live terminal dashboard showing real-time token usage, context window consumption, and request details. It intercepts LLM API traffic automatically with zero configuration, supports multiple providers (Anthropic, OpenAI, Gemini, MiniMax), and saves every prompt as markdown and JSON for debugging. The color-coded fuel gauge visualizes context usage against the user-defined limit, and a session summary shows total token usage upon exit. The tool is installed via pip and requires Python 3.10+.

Key Features

Real-time token tracking with color-coded fuel gauge (green, yellow, red) showing context usage against the limit

Prompt archive automatically saves every intercepted request as markdown and JSON for debugging

Session summary displays total tokens and request count upon exit

Zero-configuration proxy – install via pip and run commands like 'tokentap claude' or 'tokentap run'

Supports multiple LLM providers: Anthropic, OpenAI, Gemini, and MiniMax

Live terminal dashboard showing time, provider, model, token count, and latest prompt snippet

Pros & Cons

Pros

Real-time visibility into token and context usage with a visual dashboard
Automatic prompt archiving in human-readable markdown and raw JSON
Zero configuration – just install and use, no certificates or setup required
Supports multiple popular LLM providers out of the box
Session summary provides total usage metrics for cost tracking

Cons

Requires Python 3.10 or higher
Limited to CLI-based tools and supported providers (Anthropic, OpenAI, Gemini, MiniMax)
Terminal-only interface; no web or mobile dashboard
May need manual configuration for non-listed LLM tools using 'tokentap run'

Best For

Monitor and debug prompt token usage during AI application developmentTrack LLM API costs across development sessions in real-timeVisualize context window consumption to stay within model limitsReview and analyze past prompts and API request bodies for optimization

FAQ

How do I install Tokentap?

Tokentap can be installed via pip: pip install tokentap. Alternatively, you can clone the repository and install from source.

How do I start using Tokentap?

In one terminal, run 'tokentap start' to launch the dashboard. In another terminal, run a command like 'tokentap claude' or 'tokentap run --provider openai python my_app.py' to run your LLM tool through the proxy.

Which LLM providers are supported?

Tokentap supports Anthropic, OpenAI, Gemini, and MiniMax providers. Use the --provider flag with 'tokentap run' to specify the provider.

Does Tokentap save my prompts?

Yes, every intercepted request is automatically saved both as markdown (human-readable with metadata) and as JSON (raw API request body) to your chosen directory.

What information appears on the dashboard?

The dashboard shows a color-coded context usage bar, a table with time, provider, model, and token count per request, and the last prompt snippet.