> ## Documentation Index > Fetch the complete documentation index at: https://docs.veri.studio/llms.txt > Use this file to discover all available pages before exploring further. # Reserve the next item > Returns the caller's own live reservation in this queue if they hold one, else reserves the oldest pending item (or an expired reservation) for the queue's reservation_minutes; two reviewers never hold the same item. The response carries what a reviewer needs: the item, the queue instructions, the pinned evaluator specs, and the context (trace items: request metadata, the trace body and transcript; thread items: the thread's request ids and the latest trace's transcript; experiment items: input, output, expected and transcript). 204 when nothing is left to review. ## OpenAPI ````yaml /api-reference/openapi.json post /v1/annotation-queues/{id_or_name}/items/next openapi: 3.1.0 info: title: Veri API description: >- REST API for the Veri RL post-training platform. All requests require a Bearer API key (`vk_` prefix). license: name: '' version: 0.1.0 servers: - url: https://api.veri.studio description: Production security: [] tags: - name: Training jobs description: Create, monitor, and manage training jobs. - name: Datasets description: Upload and connect training datasets. - name: Deployments description: Serve trained models and run inference. - name: Volumes description: Persistent file storage mounted into jobs. - name: Models description: Custom model registry deployments serve from. - name: Regions description: Discover available launch regions. - name: GPU description: Live GPU availability by provider and region. - name: Code artifacts description: Upload custom training script bundles. - name: Billing description: Credit balance and transaction history. - name: API keys description: Create and revoke API keys. - name: Account description: The authenticated caller's identity. - name: Settings description: Account-level integrations (Weights & Biases). - name: Metrics description: Prometheus metrics export for your own observability stack. - name: Evaluators description: 'Evaluators: versioned scoring rules (LLM judge, code, human).' - name: Experiments description: >- Experiments: offline runs of a pinned dataset snapshot through a target, scored by pinned evaluators. - name: Annotation queues description: >- Human review: queue traces, threads and experiment items, reserve one at a time, score them and feed corrections back into datasets. - name: Monitors description: >- Monitors: evaluators scoring a sampled share of a deployment's live traffic. - name: Observability description: Agent conversations, agents and their traffic. - name: Public runs description: Unauthenticated reads of runs their owners published to Explore. paths: /v1/annotation-queues/{id_or_name}/items/next: post: tags: - Annotation queues summary: Reserve the next item description: >- Returns the caller's own live reservation in this queue if they hold one, else reserves the oldest pending item (or an expired reservation) for the queue's reservation_minutes; two reviewers never hold the same item. The response carries what a reviewer needs: the item, the queue instructions, the pinned evaluator specs, and the context (trace items: request metadata, the trace body and transcript; thread items: the thread's request ids and the latest trace's transcript; experiment items: input, output, expected and transcript). 204 when nothing is left to review. operationId: next parameters: - name: id_or_name in: path description: Queue id (aq_...) or name required: true schema: type: string responses: '200': description: The reserved item and its review context content: application/json: schema: $ref: '#/components/schemas/AnnotationTaskResponse' '204': description: No pending items '400': description: Evals disabled '404': description: No such queue in the caller's workspace '409': description: The queue is archived security: - bearerAuth: [] components: schemas: AnnotationTaskResponse: type: object description: A reserved item and everything needed to review it. required: - object - item - queue - evaluators - context properties: object: type: string description: Always "annotation_task". item: $ref: '#/components/schemas/AnnotationItemResponse' queue: $ref: '#/components/schemas/QueueSummary' evaluators: type: array items: $ref: '#/components/schemas/QueueEvaluator' description: Score each of these (POST .../complete takes one score per entry). context: $ref: '#/components/schemas/ReviewContext' AnnotationItemResponse: type: object required: - object - id - queue_id - target_type - status - created_at properties: object: type: string description: Always "annotation_item". id: type: string description: '`ai_...`' queue_id: type: string target_type: type: string description: trace | thread | experiment_item request_id: type: - string - 'null' description: 'Trace items: the request (trace id `veri-`).' deployment_id: type: - string - 'null' description: Trace and thread items. thread_id: type: - string - 'null' description: 'Thread items: the thread and the last request under review.' target_revision: type: - string - 'null' experiment_id: type: - string - 'null' description: Experiment-item items. item_id: type: - string - 'null' status: type: string description: |- pending | reserved | completed | skipped. An expired reservation is pending. reserved_by: type: - string - 'null' description: 'While reserved: who holds it and until when.' reserved_until: type: - string - 'null' format: date-time completed_by: type: - string - 'null' description: Who closed the item (completed or skipped) and when. completed_at: type: - string - 'null' format: date-time correction: description: The reviewer's corrected answer (any JSON value). note: type: - string - 'null' created_at: type: string format: date-time QueueSummary: type: object description: The queue facts a reviewer needs next to an item. required: - id - name - target_type - reservation_minutes properties: id: type: string name: type: string instructions: type: - string - 'null' target_type: type: string reservation_minutes: type: integer format: int32 QueueEvaluator: type: object description: A pinned evaluator version, as a reviewer scores it. required: - evaluator_id - name - version - method - scope - output_type - output_config properties: evaluator_id: type: string name: type: string description: type: - string - 'null' version: type: integer format: int32 method: type: string description: llm_judge | code | human scope: type: string description: single_turn | thread output_type: type: string description: boolean | categorical | numeric output_config: type: object description: >- boolean {pass_when}, categorical {labels, pass_labels}, numeric {min, max, pass_threshold, higher_is_better}: the values a score may take. instructions: type: - string - 'null' description: 'Human evaluators: the version''s reviewer instructions.' ReviewContext: type: object description: |- Everything a reviewer needs to judge one item. Trace items carry `request`, `trace` and `transcript`; thread items add the thread's `request_ids` (the trace and request are the revision's, whose messages hold the whole conversation); experiment items carry `experiment_item`. properties: request: oneOf: - type: 'null' - $ref: '#/components/schemas/RequestInfo' request_ids: type: - array - 'null' items: type: string description: |- Thread items: the thread's request ids, oldest first, through the revision. trace: type: - object - 'null' description: |- The Langfuse trace, bodies included; null when unavailable (see trace_error). trace_error: type: - string - 'null' description: |- Why `trace` is null: traces off, the deployment records no bodies, retention purged it, or the trace store failed. transcript: type: - array - 'null' items: type: object description: |- The conversation as chat messages: the trace's request messages plus its reply, or the experiment item's transcript. experiment_item: oneOf: - type: 'null' - $ref: '#/components/schemas/ExperimentItemInfo' RequestInfo: type: object description: A traced request's metadata. required: - id - deployment_id - created_at - status_code - latency_ms - prompt_tokens - completion_tokens properties: id: type: string deployment_id: type: string created_at: type: string format: date-time model_version_id: type: - string - 'null' thread_id: type: - string - 'null' status_code: type: integer format: int32 latency_ms: type: number format: double prompt_tokens: type: integer format: int32 completion_tokens: type: integer format: int32 finish_reason: type: - string - 'null' ExperimentItemInfo: type: object description: An experiment item as a reviewer sees it. required: - experiment_id - experiment_name - row_id - trial - kind - status properties: experiment_id: type: string experiment_name: type: string row_id: type: integer format: int64 trial: type: integer format: int32 turn: type: - integer - 'null' format: int32 description: The turn under review (per-turn items of a replayed thread). kind: type: string description: single_turn | thread turns: type: - integer - 'null' format: int32 input: {} output: {} expected: {} metadata: {} transcript: {} status: type: string description: pending | running | done | error securitySchemes: bearerAuth: type: http scheme: bearer bearerFormat: API key description: API key with the `vk_` prefix. Create one from the dashboard. ```` This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.