jents 0.1.0

Jents Capture — one line to capture your LLM usage (cost, tokens, latency) into Jents.

jents

One line to capture your LLM usage — cost, tokens, latency — into Jents.

pip install jents

import jents
jents.init(api_key="<your ingest key>", team="<your team>")

That's it. Your existing OpenAI / Anthropic calls are now captured automatically and sent to your Jents workspace — attributed to the right team (and person, if you set user_id). Only metadata is sent (model, tokens, cost, latency); your prompts and responses never leave your process.

Get an ingest key in Jents: Observability → Capture → Generate ingest key.

Options

jents.init(
    api_key="<your ingest key>",   # required
    team="Sales",                  # per-team cost rollups
    user_id="[e-mail]",   # optional per-person attribution
    app_name="my-service",         # optional service label
)

Capture Anthropic calls too:

pip install "jents[anthropic]"

Self-hosting Jents? Point at your own endpoint:

jents.init(api_key="...", endpoint="https://jents.your-company.com/api/capture")

How it works

jents is a thin convenience layer over the OpenTelemetry open standard: it configures an OTLP exporter aimed at your Jents Capture endpoint and turns on auto-instrumentation for your LLM clients. It composes cleanly with any other OpenTelemetry tracing you already run.