> ## Documentation Index
> Fetch the complete documentation index at: https://docs.omnia-voice.com/llms.txt
> Use this file to discover all available pages before exploring further.

# OTLP trace ingest

> Accepts standard OpenTelemetry traces (OTLP/HTTP, protobuf or JSON; 4 MB batch cap) at the GATEWAY host: `https://gateway.omnia-voice.com/v1/traces`. Point any OTel exporter here with `Authorization: Bearer <key>`.

Model-call spans are recognized in the official `gen_ai.*` conventions plus the Traceloop, Vercel `ai.*`, and OpenInference dialects, normalized to one shape, and stored as gradeable log entries when request logging is enabled; all other spans (and failed model calls) are kept as trace structure. Ingest is idempotent under exporter retries and rate-limited in its own bucket, separate from inference.

See the [OTLP trace ingest reference](/reference/otlp-ingest) for the full contract.



## OpenAPI

````yaml /openapi.json post /v1/traces
openapi: 3.1.0
info:
  title: Omnia Management API
  description: >-
    The management API behind the improvement loop: capture and setup, request
    logs and datasets, grades (labels), judges (criteria), evals and deploy
    gates, fine-tuning and reinforcement learning, dedicated GPU endpoints, and
    model aliases and versions. Authenticated with a workspace API key
    (sk_sovereign_...). The inference API (chat, embeddings, rerank, responses)
    is OpenAI-compatible and documented separately.


    Responses are snake_case, list endpoints on the loop products use the
    {"object": "list", "data": [...]} envelope, and refusals use the same nested
    error shape the gateway emits: {"error": {"message", "type", "code"}}.
    Request bodies on the loop products (logs, labels, criteria, evals,
    datasets, aliases) are snake_case; the training and infrastructure products
    (fine-tuning, GRPO, environment tools, dedicated, model-version adoption)
    validate camelCase bodies, and each schema below says which it is. Endpoints
    that spend money require a key minted by a workspace owner or admin and
    return 403 otherwise.
  version: 1.0.0
servers:
  - url: https://gateway.omnia-voice.com
    description: Production
  - url: https://platform.omnia-voice.com/api
    description: Production (legacy alias — same API, older base URL)
security:
  - bearerAuth: []
paths:
  /v1/traces:
    post:
      summary: OTLP trace ingest
      description: >-
        Accepts standard OpenTelemetry traces (OTLP/HTTP, protobuf or JSON; 4 MB
        batch cap) at the GATEWAY host:
        `https://gateway.omnia-voice.com/v1/traces`. Point any OTel exporter
        here with `Authorization: Bearer <key>`.


        Model-call spans are recognized in the official `gen_ai.*` conventions
        plus the Traceloop, Vercel `ai.*`, and OpenInference dialects,
        normalized to one shape, and stored as gradeable log entries when
        request logging is enabled; all other spans (and failed model calls) are
        kept as trace structure. Ingest is idempotent under exporter retries and
        rate-limited in its own bucket, separate from inference.


        See the [OTLP trace ingest reference](/reference/otlp-ingest) for the
        full contract.
      operationId: ingestTraces
      requestBody:
        required: true
        content:
          application/x-protobuf:
            schema:
              type: string
              format: binary
          application/json:
            schema:
              type: object
              description: OTLP/JSON ExportTraceServiceRequest.
      responses:
        '200':
          description: Per-batch accounting.
          content:
            application/json:
              schema:
                type: object
                properties:
                  accepted:
                    type: integer
                    description: Gradeable log rows written.
                  duplicate:
                    type: integer
                    description: Spans dropped as exporter retries.
                  spans_stored:
                    type: integer
                    description: Structure rows written.
                  skipped_not_gen_ai:
                    type: integer
                  skipped_no_content:
                    type: integer
                    description: Spans unstorable in any form.
                  content_logging:
                    type: string
                    enum:
                      - enabled
                      - disabled
                  stored_as_structure_only:
                    type: integer
                    description: >-
                      Model calls kept as structure: logging off, redacting
                      exporter, or a failed call.
                  unrecognized_attr_namespaces:
                    type: object
                    description: >-
                      Attribute-KEY namespaces seen but not mapped (never
                      values).
        '413':
          description: Batch over 4 MB.
        '429':
          description: Ingest budget spent — telemetry cannot consume the inference budget.
components:
  securitySchemes:
    bearerAuth:
      type: http
      scheme: bearer
      description: >-
        Your workspace API key, e.g. `sk_sovereign_...`, sent as `Authorization:
        Bearer <key>`.

````