Skip to content
llmprof

Roadmap

Available today

Section titled “Available today”
  • Zero-code-change proxy for OpenAI and Anthropic APIs (streaming-safe).
  • Per-component token attribution with per-tool drill-down.
  • Cost for 100+ models, fully overridable.
  • Context flame graph, trends, context timeline, and cost leaderboard views.
  • Waste detector with a reclaimable-per-month headline.
  • Python SDK for precise, in-code component labels.
  • Pluggable storage (SQLite default; backend abstraction for a centralized database).
  • npx llmprof launcher - run the proxy with no Python install. See Installation.
  • JavaScript / TypeScript SDK - the same component-labeling API as Python. See the JS/TS SDK.
  • Hosted live demo - the real dashboard on a recorded session, no install. Try it in your browser.

Next

Section titled “Next”
  • Postgres backend - a shared/team dashboard on the existing store abstraction.

Not in scope

Section titled “Not in scope”

llmprof is a focused profiler, not a full observability platform. It does not do eval suites, prompt management, or model hosting - it diagnoses where your tokens and dollars go and tells you what to cut.