> ## 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. # Create a dataset from logs > Curate logged traffic into a managed training dataset. Curation is automatic: errored, truncated (`finish_reason=length`), empty, and duplicate exchanges are dropped, and every drop is reported by reason. With `holdout_pct`, a second **disjoint** `{name}-eval` dataset is created — no example appears in both, so it's a valid eval set rather than the model's own homework. Pass its id as `sample_filters.dataset_id` when queueing an eval. Requires an owner/admin key. ## OpenAPI ````yaml /openapi.json post /v1/datasets/from_logs 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/datasets/from_logs: post: tags: - Logs summary: Create a dataset from logs description: >- Curate logged traffic into a managed training dataset. Curation is automatic: errored, truncated (`finish_reason=length`), empty, and duplicate exchanges are dropped, and every drop is reported by reason. With `holdout_pct`, a second **disjoint** `{name}-eval` dataset is created — no example appears in both, so it's a valid eval set rather than the model's own homework. Pass its id as `sample_filters.dataset_id` when queueing an eval. Requires an owner/admin key. operationId: createDatasetFromLogs requestBody: required: true content: application/json: schema: $ref: '#/components/schemas/DatasetFromLogs' responses: '201': description: The created dataset(s) and the curation summary. content: application/json: schema: $ref: '#/components/schemas/DatasetFromLogsResult' '400': $ref: '#/components/responses/BadRequest' '401': $ref: '#/components/responses/Unauthorized' '403': $ref: '#/components/responses/Forbidden' components: schemas: DatasetFromLogs: type: object required: - name properties: name: type: string maxLength: 80 example: support-bot-v1 holdout_pct: type: integer minimum: 0 maximum: 50 default: 0 description: >- Also create a second, DISJOINT "{name}-eval" dataset holding this percentage of examples. No example appears in both — that's what makes it valid as an eval source. example: 10 filters: type: object description: >- Which logged exchanges to curate. Omit for all of them. Keys: model, tag (the task label sent as X-Omnia-Tag; a dataset scoped to one task beats one built from mixed traffic), segment (an auto-detected traffic segment, a `segment` value from GET /v1/logs), finish_reason, cache_hit (boolean), start and end (unix seconds). properties: model: type: string tag: type: string segment: type: string description: >- Curate only one auto-detected traffic segment (a `segment` value from GET /v1/logs) — one application surface, even when nothing was tagged. finish_reason: type: string cache_hit: type: boolean start: type: integer description: Unix seconds, inclusive. end: type: integer description: Unix seconds, exclusive. sources: type: object description: >- Which models feed the build. The same task (tag) usually spans every model you route it to; this chooses among them. properties: models: type: array description: >- Include-list: only these models' exchanges feed the build. An empty list means no restriction. Rejected with 400 unless every entry is a non-empty string. items: type: string quality: type: object description: >- The quality ladder (default cleaned). In every mode, exchanges a human graded FAIL never enter a dataset. properties: mode: type: string description: >- cleaned: mechanical curation only. graded: keep only exchanges a human graded pass (free). judge: a calibrated judge keeps only passing conversations; judge calls are billed as usage and refused up front (402) if the wallet can't cover a conservative estimate. enum: - cleaned - graded - judge criterion_id: type: string description: >- Required for mode "judge". The criterion must be calibrated (aligned tier), not drift-flagged, and request-unit; a segment-bound judge is only usable on builds scoped to its segment. DatasetFromLogsResult: type: object properties: training_name: type: string training_count: type: integer eval_name: type: string description: Present when holdout_pct > 0. eval_count: type: integer summary: type: object description: >- Curation accounting: total, kept, dropped {unparseable, errored, truncated, empty, duplicate, human_failed}, and the multi-turn folding counts. Every drop is counted by reason, so a dataset is never silently thinner than you think. properties: total: type: integer kept: type: integer dropped: type: object properties: unparseable: type: integer errored: type: integer truncated: type: integer description: >- finish_reason=length — cut off mid-answer, misleading as training data. empty: type: integer duplicate: type: integer description: >- Byte-identical exchange already kept (cache-hit replays collapse here). quality: type: object description: >- What the quality gate actually did: mode, criterion_id and criterion_name (judge mode), the judge's kappa at build time, human_pass_kept, ungraded_excluded (graded mode), judged / judge_passed / judge_failed / judge_unparsed and judge_spend_usd (judge mode). A dataset never claims more than what happened. Error: type: object properties: error: type: object properties: message: type: string type: type: string description: >- invalid_request_error, insufficient_quota, rate_limit_error, or api_error. code: type: string description: >- Machine-stable cause, e.g. invalid_api_key, not_found, insufficient_permissions, precondition_failed. required: - message - type - code description: >- Every refusal — gateway and management API alike — uses this one envelope. responses: BadRequest: description: Malformed request or invalid field. content: application/json: schema: $ref: '#/components/schemas/Error' Unauthorized: description: Missing, malformed, or revoked API key. content: application/json: schema: $ref: '#/components/schemas/Error' example: error: message: Invalid API key type: invalid_request_error code: invalid_api_key Forbidden: description: The key lacks the required owner/admin permission. content: application/json: schema: $ref: '#/components/schemas/Error' securitySchemes: bearerAuth: type: http scheme: bearer description: >- Your workspace API key, e.g. `sk_sovereign_...`, sent as `Authorization: Bearer `. ````