tokenshark-bin 0.1.0

Patch your LLM client with one import and see real-time token costs in your terminal.

TokenShark

Patch your LLM client with one import and see real-time token costs in your terminal.

import tokenshark
tokenshark.monitor()
# OpenAI and Anthropic calls are now logged with cost, latency, and
# (optionally) your own metadata tags.

Status

Early build — core patching + SQLite cost tracking (proxy.py, tracker.py, config.py, exceptions.py) is in place. cli.py, dashboard.py, alerts.py, and tests are still to come.

Install (once published)

pip install tokenshark[openai,anthropic]

---

This is a placeholder — pyproject.toml's readme field points here, and hatchling needs the file to exist to build package metadata at all, so pip install -e . would fail on a fresh clone without it. Not one of the 5 requested files — added purely to keep today's pip install -e . from breaking. Replace with a real README on Day 3 polish.