Which reference inputs do current video-generation vendor docs accept for keeping a character or subject the same across separate generations, and what limits and failure cases do they state?

Vendor docs read on 2026-10-22 offer three different mechanisms, not one: (a) reference images that condition a generation (Google Veo 3.1: up to 3 images of a person, character or product, and 8 s output only; Runway's hosted third-party models: e.g. up to 7 for Kling O3, 10 images plus 5 videos plus 5 audio for WAN 3.0), (b) first/last or multi-position keyframes (Luma Ray 3.2: 1-64 keyframes, exclusive with start/end frame), and (c) a reusable character asset (OpenAI Sora, non-human only, name must be in the prompt; the Sora API was shut down on 2026-09-24). None of the pages read publishes a measurement of cross-shot identity; Runway's own Gen-4 / Gen-4.5 image-to-video entries list only a single prompt image in the docs read. Vendor statements are claims; Kling's docs could not be read. Curator correction 2026-10-24 (curator source re-verification record CUR-0618-02; sources re-fetched 2026-10-24, excerpts re-located by curator GET): The Veo 3.1 reference-image limits above were read from the Gemini API Veo documentation. The Gemini API Veo 3.1 Preview models were shut down on October 22, 2026 (warning on https://ai.google.dev/gemini-api/docs/veo.md.txt, curator GET 2026-10-24T03:16:31Z, sha256 6ddf6b9ca73d4cb3fb6bc28233b13a15e009ae645fc2023a2956856139ef6917); Google points users to gemini-omni-1.1-flash or GA Veo models on the Gemini Enterprise Agent Platform. Check the reference-image limits on the Agent Platform model page before relying on them; the Gemini API page's code samples still use the shut-down preview id. Curator qualifications: All excerpts located; three vendor mechanisms (reference images, character assets, keyframes) dated 2026-10-22. Evidence status: statements located in the cited sources by a correlated desk review (Claude); not executed; not independently reviewed.

{
  "citations": [
    {
      "url": "https://ai.google.dev/gemini-api/docs/veo.md.txt",
      "note": "candidate evidence excerpts"
    },
    {
      "url": "https://ai.google.dev/gemini-api/docs/video.md.txt",
      "note": "candidate evidence excerpts"
    },
    {
      "url": "https://docs.cloud.google.com/gemini-enterprise-agent-platform/models/video/generate-videos-from-references",
      "note": "candidate evidence excerpts"
    },
    {
      "url": "https://ai.google.dev/gemini-api/docs/omni.md.txt",
      "note": "candidate evidence excerpts"
    },
    {
      "url": "https://docs.dev.runwayml.com/assets/inputs/",
      "note": "candidate evidence excerpts"
    },
    {
      "url": "https://docs.dev.runwayml.com/recipes/reference-media/",
      "note": "candidate evidence excerpts"
    },
    {
      "url": "https://docs.dev.runwayml.com/recipes/multi-shot-video/",
      "note": "candidate evidence excerpts"
    },
    {
      "url": "https://developers.openai.com/api/docs/guides/video-generation.md",
      "note": "candidate evidence excerpts"
    },
    {
      "url": "https://docs.agents.lumalabs.ai/guides/videos/generation/index.md",
      "note": "candidate evidence excerpts"
    }
  ]
}