TraptureIQ
Sign inStart free
Documentation

TraptureIQ docs

See what your Google ADK agents cost, what they did, and whether they are any good.

What TraptureIQ does

You point a Google ADK agent at TraptureIQ and it starts recording every LLM call. From that one stream you get four things in the console:

ModuleAnswers
CostWhat am I spending, on which agent, which model, which user?
ObservabilityWhat did the agent actually do, and where did it fail?
EvalIs it any good, and does it hold up under load?
ToolsPrompts, parameters and token maths, versioned, outside your code.
Google ADK only, for nowTraptureIQ currently supports agents built with Google ADK. That is a deliberate scope, not a limitation we are hiding: ADK emits the telemetry this product is built on, which is why connecting takes two environment variables and no code change. Other frameworks are on the roadmap. Both Cloud Run and Agent Engine are verified end to end, with no edit to your agent in either case. Agent Engine requires ADK ≥ 2.11.0.

Connecting takes two environment variables

ADK already emits OpenTelemetry spans carrying token usage. TraptureIQ reads them, so there is nothing to add to agent.py, no SDK, no plugin, no agent registration.

Two are required: the traces endpoint and the API key header. The others are recommended: the logs endpoint, the agent's name, and the environment that keeps local runs apart from production.

export OTEL_EXPORTER_OTLP_TRACES_ENDPOINT="https://api.traptureiq.com/v1/traces"
export OTEL_EXPORTER_OTLP_LOGS_ENDPOINT="https://api.traptureiq.com/v1/logs"
export OTEL_EXPORTER_OTLP_HEADERS="X-API-Key=tiq_your_key_here"
export OTEL_SERVICE_NAME="my-agent"
export OTEL_RESOURCE_ATTRIBUTES="deployment.environment=local"

Start here