Integrations
It grades the endpoint A–F and writes a self-contained report. It works with OpenAI-compatible servers (vLLM, TGI, Ollama, llama.cpp and others) and the OpenAI and Anthropic APIs.
CI
The GitHub Action fails a pull request whose latency or cost regressed against a baseline report.
Kubernetes
A Helm chart scans an in-cluster endpoint on a schedule and keeps the latest report.
Coding agents
tokvera mcp is an MCP server for coding agents. It sends load only to an endpoint you named, and only after you agreed to it.
Its tools: scan_demo, scan_plan, scan (needs confirm: true), summarize_report, verify_report, list_suites and doctor.
Usage logs
tokvera capture is a small reverse proxy for OpenAI-compatible chat completions that writes one usage-log line per call, for clients with no OpenTelemetry instrumentation.
Audit kit
The audit kit adds OpenTelemetry GenAI instrumentation for Go, Python and Node (tokvera instrument), a local observability stack (tokvera stack up) and a caching proxy (tokvera cache).