Skip to main content
terminal
pip install botgauge installs the tracing SDK and the botgauge command. It depends only on OpenTelemetry. On its own it records the turns, agents, and spans your own code creates. It does not install tracing for every framework BotGauge supports, so you add only what your app uses. The package requires Python 3.10 or newer.

​
Framework extras

Each extra adds the instrumentation that records one framework’s LLM calls (model, prompts, token counts) and tool calls. botgauge setup --install finds the frameworks installed in your environment without importing them, then installs the BotGauge extra for each one at your installed BotGauge version. It uses pip, or uv in an environment created by uv venv. To preview without installing anything, run botgauge setup --detect. You can also name an extra yourself, for example in requirements.txt. This installs LangChain tracing with the versions BotGauge verified.
terminal
You can combine extras in one install, for example pip install "botgauge[mcp,openai]". The supported frameworks page lists the framework versions BotGauge verified with each extra.

​
Verify the installation

terminal
botgauge doctor prints a read-only report of the current process: the tracer provider, active instrumentors, and any warnings. Add --json for machine-readable output. See the CLI reference for every command.

Install the SDK from your project

Start in BotGauge to get your API key and then install the package with pip.

Try BotGauge →

​
Next step

Manual tracing shows how to wrap turns and agents in code. If you’d rather not edit your app, go to zero-code instrumentation.