An open governance specification for AI workloads, compatible with OpenTelemetry.

OTel-native typed primitives for LLM cost attribution and telemetry — published on PyPI.

OTel Collector processor that normalizes vendor/framework LLM attributes to canonical gen_ai.* semantic conventions. One dashboard to rule them all.

OpenTelemetry GenAI semantic conventions for Go: drop-in HTTP middleware, gRPC interceptors, and client transports that emit spec-aligned inference + tool spans with multi-turn conversation correlation.

Validate OpenTelemetry GenAI spans against the official semantic conventions — 10 rules covering required attrs (provider/operation), token usage names + ranges, legacy attr-name detection, span kind, name format. Library + CLI.

OpenTelemetry semantic conventions and instrumentation for agent provenance, derivation lineage, and acceptance criteria evaluation. Fills the Microsoft AI stack observability gap.

Curated OTel GenAI conformance corpus — categorized OTLP/JSON fixtures (valid / dirty / shape) with declarative manifest. Drop-in regression target for any GenAI consumer: validators, exporters, rollups, dashboards.

Reference implementation of OpenTelemetry GenAI semantic conventions — wrappers for Anthropic, OpenAI, Vertex AI, and Bedrock that emit spec-compliant spans. Dashboards for Phoenix, Langfuse, and Cloud Trace.

OpenTelemetry instrumentation package for the Dust SDK. Automatically instruments agent interactions using OTel GenAI Semantic Conventions.

Workflow-neutral work ledger schemas, semantic conventions, conformance, and receipt validation for human, AI, CI, and tool-driven work.

Λ-signed OpenTelemetry exporter for SZL audit fibers. Doctrine v6 · DOI 10.5281/zenodo.20424995