Skip to content
reaatechREAATECH

reaatech/otel-genai-semconv

0Last commit: May 2, 2026GitHub →

These packages provide instrumented wrappers for OpenAI, Anthropic, Vertex AI, and AWS Bedrock that emit OpenTelemetry-compliant spans for LLM interactions. You would adopt them to gain standardized observability into token usage, costs, streaming performance, and error rates across multiple AI providers. The collection is built on a shared instrumentation framework that uses a zero-config wrapper pattern to inject lifecycle hooks, circuit breakers, and PII redaction directly into existing SDK clients.

Packages

9 packages

@reaatech/otel-genai-semconv-anthropic

v0.1.0
Instruments the Anthropic Node.js SDK to automatically emit OpenTelemetry spans compliant with GenAI semantic conventions. It provides a class that wraps the Anthropic client to capture request metadata, token usage, and streaming metrics.
status
published
published
10 hours ago

@reaatech/otel-genai-semconv-bedrock

v0.1.0
Instruments the AWS Bedrock Runtime SDK to automatically emit OpenTelemetry spans compliant with GenAI semantic conventions. It provides an `instrument` method that wraps the `BedrockRuntimeClient.send` function to capture request and response metadata for supported model families.
status
published
published
10 hours ago

@reaatech/otel-genai-semconv-core

v0.1.0
Provides TypeScript types, Zod schemas, and a span-builder utility for implementing OpenTelemetry GenAI semantic conventions. It exports constants for attributes, events, and metrics to ensure OTel-compliant telemetry across LLM operations.
status
published
published
10 hours ago

@reaatech/otel-genai-semconv-exporters

v0.1.0
Provides OpenTelemetry `SpanExporter` classes that filter for `gen_ai.*` spans and transform them into the native formats required by Arize Phoenix, Langfuse, and Google Cloud Trace. These exporters implement the standard OTel `SpanExporter` interface to enable direct integration with existing OpenTelemetry SDK pipelines.
status
published
published
10 hours ago

@reaatech/otel-genai-semconv-instrumentation

v0.1.0
Provides a suite of classes for building LLM observability instrumentation, including utilities for OpenTelemetry tracer management, streaming response aggregation, error classification, and resilience patterns like retries and circuit breaking. It serves as the base framework for provider-specific instrumentation packages.
status
published
published
10 hours ago

@reaatech/otel-genai-semconv-observability

v0.1.0
Provides pre-configured OpenTelemetry SDK initialization, Pino-based structured logging, and GenAI-specific metrics instrumentation for LLM applications. It exports utility functions to bootstrap observability, manage trace exporters, and expose health check endpoints.
status
published
published
10 hours ago

@reaatech/otel-genai-semconv-openai

v0.1.0
Instruments the OpenAI Node.js SDK to automatically emit OpenTelemetry spans compliant with GenAI semantic conventions. It provides an `OpenAIInstrumentation` class that wraps the client's `chat.completions.create` method to capture request metadata, token usage, and cost metrics.
status
published
published
10 hours ago

@reaatech/otel-genai-semconv-utils

v0.1.0
Provides classes for estimating LLM token usage, calculating request costs based on provider-specific pricing, and recursively redacting PII from nested objects. It exposes `TokenCounter`, `CostCalculator`, and `PIIRedactor` classes that integrate with common LLM providers like OpenAI, Anthropic, and Bedrock.
status
published
published
10 hours ago

@reaatech/otel-genai-semconv-vertexai

v0.1.0
Instruments the Google Generative Language (Vertex AI) SDK to automatically emit OpenTelemetry spans following GenAI semantic conventions. It provides an `instrument` function that wraps the `generateContent` method to capture request metadata, generation configurations, and response events.
status
published
published
10 hours ago

Comments

Sign in with GitHub to comment and vote.

Loading comments…