---
title: "Get an inference record"
method: GET
path: "/inferences/{inference_id}"
tags: ["inference-history"]
---

# Get an inference record

`GET /inferences/{inference_id}`

Retrieve a single inference record by ID. Includes LLM-as-Judge verdict and score when judging has completed (fields are null until then).

## Path parameters

- `inference_id` string, required

## Response `200`

The inference record.

- InferenceRecord
  - `id` string, required
  - `user_id` string, required
  - `model_id` string, required
  - `model_name` string
  - `task` string
  - `input` string, required
  - `output` unknown
  - `latency_ms` integer
  - `tokens` integer
  - `input_tokens` integer
  - `output_tokens` integer
  - `cache_read_tokens` integer
  - `cache_write_tokens` integer
  - `source` 'api' | 'ui', required
  - `status` 'success' | 'failed', required
  - `error_type` 'validation' | 'timeout' | 'model_not_ready' | 'model_not_found' | 'model_not_supported' | 'capacity_exhausted' | 'internal'
  - `error_message` string
  - `created_at` string, date-time, required
  - `project_id` string
  - `training_job_id` string
  - `provider` string
  - `base_model` string
  - `metadata` object
  - `human_verdict` 'correct' | 'incorrect'
  - `human_corrected_output` unknown
  - `human_feedback_notes` string
  - `human_feedback_at` string, date-time
  - `llmaj_verdict` 'pass' | 'fail' | 'uncertain' — LLM-as-Judge verdict. Null until judging completes.
  - `llmaj_score` number — LLM-as-Judge confidence score [0.0, 1.0]. Null until judging completes.
  - `llmaj_judged_at` string, date-time
  - `llmaj_reasoning` string

## Other responses

- `401` — Missing or invalid API key.
- `404` — Resource not found.

---

[API](https://skmtc.net/fastino-ai/apis/pioneer-inference-api.md) · [All operations](https://skmtc.net/fastino-ai/apis/pioneer-inference-api/llms.txt) · [OpenAPI document](https://skmtc-service-staging.skmtc.workers.dev/v1/apis/fastino-ai/pioneer-inference-api/revisions/2541b0de636c/schema)