{
  "id": "whisper-1",
  "identity": {
    "family": "whisper",
    "display_name": "OpenAI Whisper API (whisper-1)",
    "hub_id": null,
    "api_name": "whisper-1",
    "revision": null,
    "variants": [],
    "license": "OpenAI API terms of use",
    "commercial_use": true,
    "commercial_use_note": "Paid OpenAI API under OpenAI's terms.",
    "release_date": "2023-03",
    "paper_url": "https://arxiv.org/abs/2212.04356",
    "repo_url": null
  },
  "task": {
    "category": "transcription",
    "modalities_in": [
      "audio",
      "video"
    ],
    "modalities_out": [
      "text"
    ],
    "best_at": "Speech-to-text for video and audio with no GPU in the cluster; the production default for multimodal_extractor transcription.",
    "worst_at": "Provider rate limits (a Primis batch lost its transcripts to a 429), per-minute cost, and audio sent off-box."
  },
  "io": {
    "input": {
      "audio": "audio chunks from AudioParallelChunking",
      "language": "optional forced language"
    },
    "output": {
      "text": "transcript, or verbose_json segments when segment timestamps are wanted"
    }
  },
  "mixpeek_wiring": {
    "extractors": [
      {
        "extractor": "multimodal_extractor@v1",
        "selected_by": "run_transcription=true and use_whisper_api (HARDWARE_WHISPER_USE_API=true, or forced when the cluster has no GPU)"
      },
      {
        "extractor": "multimodal_extractor@v2",
        "selected_by": "run_transcription=true and use_whisper_api (HARDWARE_WHISPER_USE_API=true, or forced when the cluster has no GPU)"
      }
    ],
    "feature_uris": [],
    "vector_indexes": [],
    "retriever_stages": [],
    "composes_with": [
      "video -> audio chunks -> whisper-1 transcript -> E5 transcription vector"
    ]
  },
  "run": {
    "status": "api",
    "status_reason": "OpenAI audio transcriptions API, called from WhisperTranscriptionBatch; universal__api also maps whisper-1 for POST /v1/inference.",
    "serve_apps": [
      "universal__api"
    ],
    "runtime": "OpenAI audio.transcriptions.create (engine.inference.openai.whisper_api.services)",
    "weights_format": "api",
    "trust_remote_code": false,
    "size_on_disk_mb": null,
    "params_millions": null,
    "preprocessing": "ffmpeg audio extraction and chunking per segment.",
    "postprocessing": "Text, or verbose_json segments mapped to timestamps.",
    "serving_shape": "both",
    "batch_size": null,
    "serve_batch": null,
    "scale_to_zero": {
      "min_replicas": 0,
      "max_replicas": 6,
      "idle_timeout_s": 600,
      "memory_gb": null,
      "num_cpus": 0.1,
      "num_gpus": 0,
      "cold_start_s": {
        "warm_node_s": null,
        "new_node_s": null,
        "source": null
      },
      "cpu_fallback": true
    }
  },
  "cost": {
    "unmeasured": true,
    "measured": null,
    "note": "Billed per audio minute by OpenAI; no committed measurement."
  },
  "evidence": {
    "unmeasured": true,
    "items": []
  },
  "refs": [
    "api://openai/whisper-1"
  ],
  "formats": [
    {
      "ref": "api://openai/whisper-1",
      "format": "api",
      "file": null,
      "loader": "provider SDK, called by model name",
      "supported_today": true
    }
  ],
  "use_cases": [
    "transcripts"
  ],
  "recommended": {
    "value": false,
    "for": [],
    "reason": null
  },
  "demos": [],
  "seo": {
    "title": "OpenAI Whisper API (whisper-1): speech-to-text model",
    "description": "OpenAI Whisper API (whisper-1) is a audio and video speech-to-text model (OpenAI API terms of use). Called on Mixpeek through the provider API.",
    "keywords": [
      "OpenAI Whisper API (whisper-1)",
      "whisper",
      "speech-to-text",
      "whisper-1",
      "transcripts"
    ],
    "faqs": [
      {
        "q": "What is OpenAI Whisper API (whisper-1) best at?",
        "a": "Speech-to-text for video and audio with no GPU in the cluster; the production default for multimodal_extractor transcription."
      },
      {
        "q": "What are the limits of OpenAI Whisper API (whisper-1)?",
        "a": "Provider rate limits (a Primis batch lost its transcripts to a 429), per-minute cost, and audio sent off-box."
      },
      {
        "q": "Can OpenAI Whisper API (whisper-1) be used commercially?",
        "a": "Yes. Paid OpenAI API under OpenAI's terms."
      },
      {
        "q": "How does OpenAI Whisper API (whisper-1) run on Mixpeek?",
        "a": "OpenAI audio transcriptions API, called from WhisperTranscriptionBatch; universal__api also maps whisper-1 for POST /v1/inference."
      }
    ]
  },
  "last_verified": "2026-11-05",
  "sources": [
    "https://platform.openai.com/docs/guides/speech-to-text",
    "server/engine/inference/openai/whisper_api/services.py",
    "server/engine/extractors/multimodal_extractor/v2/pipeline.py",
    "server/infra/hardware/config.py",
    "server/engine/inference/llm/routes.py",
    "customers/primis/BENCHMARK_RESULTS.md"
  ],
  "registry": {
    "universal__api": {
      "deployment": "universal__api",
      "service_id": "universal/api",
      "routes": "engine/inference/llm/routes.py",
      "route_prefix": "/universal/api",
      "enabled": true,
      "description": "Universal inference endpoint routing to multiple providers and modalities. Supports: OpenAI (GPT, embeddings, Whisper), Google (Gemini, Vertex embeddings), Anthropic (Claude). Routes based on model parameter.",
      "capabilities": {
        "category": "generation",
        "modality_input": [
          "text"
        ],
        "modality_output": "text",
        "supports_batch": true,
        "reranking_mode": null
      },
      "vector_config": null,
      "serve": {
        "min_replicas": 0,
        "max_replicas": 6,
        "idle_timeout_s": 600,
        "target_ongoing_requests": 25,
        "num_cpus": 0.1,
        "num_gpus": 0,
        "memory_gb": null,
        "pinnable": false,
        "placement": "worker",
        "enabled": true
      },
      "interactive_twin": false,
      "interactive_serve": null,
      "schemas": {
        "input": null,
        "output": null,
        "config": null
      }
    }
  },
  "templates": [],
  "usage": {
    "namespaces": null,
    "collections": null,
    "uncounted": true,
    "templates": 0,
    "counted_at": null,
    "source": null
  },
  "facets": {
    "status": "api",
    "category": "transcription",
    "family": "whisper",
    "license": "OpenAI API terms of use",
    "commercial_use": true,
    "modalities_in": [
      "audio",
      "video"
    ],
    "modalities_out": [
      "text"
    ],
    "use_cases": [
      "transcripts"
    ],
    "recommended": false,
    "release_date": "2023-03",
    "params_millions": null,
    "size_on_disk_mb": null,
    "vector_dims": null,
    "vector_distance": null,
    "weights_format": "api",
    "formats": [
      "api"
    ],
    "loadable_today": true,
    "trust_remote_code": false,
    "cpu_fallback": true,
    "num_gpus": 0,
    "memory_gb": null,
    "min_replicas": 0,
    "serving_shape": "both",
    "served": true,
    "ref_schemes": [
      "api"
    ],
    "has_evidence": false,
    "cost_measured": false,
    "template_count": 0,
    "usage_namespaces": null,
    "usage_collections": null,
    "usage_uncounted": true,
    "has_demo": false,
    "rank_score": 2
  },
  "rank_score": 2
}