Home

    ---
    title: apogee-ai-observability
    description: OpenTelemetry tracing tailored for LLM calls — token counts, model, latency, cost.
    ---

    # apogee-ai-observability

    > OpenTelemetry tracing tailored for LLM calls — token counts, model, latency, cost.

    ## What is in this package

    | Capability | What you get |

| --- | --- | | Span enrichment | trace_llm_call decorator/context manager | | Cost meter | Per-call USD estimate from token counts + model price table | | Exporters | OTLP / Jaeger / Tempo via standard OpenTelemetry SDK |

    ## Install

    ```bash
    pip install apogee-ai-observability
    ```

    ## Where it fits

    This package is part of the [Apogee Framework](/framework/) ecosystem. It can be used:

    - **Standalone** — install and use directly via `pip` (or `npm` for `@apogee/ai-ui-react`).
    - **As a plugin** — generated projects detect it and register matching CLI subcommands automatically.

    See the [framework's modules overview](/framework/modules/overview/) for how all 28 packages fit together.

    ## Read next

    - [Installation](getting-started/installation.md)
    - [Quickstart](getting-started/quickstart.md)
    - [Reference](reference/index.md)