feat: add opentelemetry instrumentation for tracing

[marcklingen]

Context

I'm one of the maintainers of github.com/langfuse/langfuse. pipecat is growing in popularity and I've gotten multiple requests by community members who'd like to get traces of their pipecat applications in order to evaluate them.

There's an open PR that adds langfuse tracing using the Langfuse Python SDK: #847

There's another PR that adds Coval: #1187

Suggestion

I'd suggest to follow the example of the AI SDK (docs) by adding native Opentelemetry-based instrumentation to pipecat. Thereby, the implementation is vendor-agnostic and teams can use an LLM-specific backend like Langfuse (tracing, evals, prompt management) or their traditional observability platforms (Grafana, Datadog, ...).

Potential addition next to traces: Exporting metrics (pipecat docs) via OpenTelemetry would allow teams to easily track them within their observability stack.

Resources

• https://opentelemetry.io/blog/2025/ai-agent-observability/
• https://opentelemetry.io/docs/specs/semconv/gen-ai/
• https://langfuse.com/docs/opentelemetry/get-started

[marcklingen]

@markbackman let me know in case you are interested to chat, happy to help explore this

[pesterhazy]

We'd love to see this as well, given that OTel is the industry standard. It would allow us to forward telemetry to any provider, including LangFuse, BetterStack, HoneyComb