End-to-End Traces
Every LLM request captured with OpenTelemetry — latency, TTFB, token counts, and cost — all in a unified trace tree with W3C Trace Context propagation.
Candela — LLM Observability Platform
Open-source observability for LLM traffic. Trace requests, monitor costs, and manage providers across your entire AI infrastructure.
What You Get
Section titled “What You Get”Multi-Provider Routing
Route to OpenAI, Gemini, Anthropic, Ollama, and LM Studio through a single endpoint. Smart routing picks the right backend automatically.
Cost Intelligence
Real-time cost calculation per request, per model, per user. Set budgets and get alerts before they’re exceeded.
Drop-In Integration
Works with Google ADK, LangChain, LiteLLM, raw HTTP — anything that talks to an LLM endpoint. Point base_url at Candela and you’re done.
Choose Your Component
Section titled “Choose Your Component” candela-local Developer proxy + runtime manager. Run local & cloud models with built-in observability.
candela-server Full backend with dashboard UI, BigQuery storage, budget enforcement, and team RBAC.
candela-sidecar Lightweight < 5MB proxy for production containers. Pub/Sub + OTLP export.
candela-desktop Flutter desktop app for provider management and trace visualization.