Video generation

OpenClaw agents can generate videos from text prompts, reference images, or existing videos. Sixteen provider backends are supported, each with different model options, input modes, and feature sets. The agent picks the right provider automatically based on your configuration and available API keys.

OpenClaw treats video generation as three runtime modes:

• generate — text-to-video requests with no reference media.
• imageToVideo — request includes one or more reference images.
• videoToVideo — request includes one or more reference videos.

Providers can support any subset of those modes. The tool validates the active mode before submission and reports supported modes in action=list.

Quick start

```bash
export GEMINI_API_KEY="your-key"
```

The agent calls `video_generate` automatically. No tool allowlisting
is needed.

How async generation works

Video generation is asynchronous. When the agent calls video_generate in a session:

1. OpenClaw submits the request to the provider and immediately returns a task id.
2. The provider processes the job in the background (typically 30 seconds to 5 minutes depending on the provider and resolution).
3. When the video is ready, OpenClaw wakes the same session with an internal completion event.
4. The agent tells the user and attaches the finished video. In group/channel chats that use message-tool-only visible delivery, the agent relays the result through the message tool instead of OpenClaw posting it directly.

While a job is in flight, duplicate video_generate calls in the same session return the current task status instead of starting another generation. Use openclaw tasks list or openclaw tasks show <taskId> to check progress from the CLI.

Outside of session-backed agent runs (for example, direct tool invocations), the tool falls back to inline generation and returns the final media path in the same turn.

Generated video files are saved under OpenClaw-managed media storage when the provider returns bytes. The default generated-video save cap follows the video media limit, and agents.defaults.mediaMaxMb raises it for larger renders. When a provider also returns a hosted output URL, OpenClaw can deliver that URL instead of failing the task if local persistence rejects an oversized file.

Task lifecycle

Check status from the CLI:

openclaw tasks list
openclaw tasks show <taskId>
openclaw tasks cancel <taskId>

If a video task is already queued or running for the current session, video_generate returns the existing task status instead of starting a new one. Use action: "status" to check explicitly without triggering a new generation.

Supported providers

Some providers accept additional or alternate API key env vars. See individual provider pages for details.

Run video_generate action=list to inspect available providers, models, and runtime modes at runtime.

Capability matrix

The explicit mode contract used by video_generate, contract tests, and the shared live sweep:

Tool parameters

Required

Content inputs

<ParamField path="image" type="string">Single reference image (path or URL).</ParamField> <ParamField path="images" type="string[]">Multiple reference images (up to 9).</ParamField> <ParamField path="imageRoles" type="string[]"> Optional per-position role hints parallel to the combined image list. Canonical values: first_frame, last_frame, reference_image. </ParamField> <ParamField path="video" type="string">Single reference video (path or URL).</ParamField> <ParamField path="videos" type="string[]">Multiple reference videos (up to 4).</ParamField> <ParamField path="videoRoles" type="string[]"> Optional per-position role hints parallel to the combined video list. Canonical value: reference_video. </ParamField> <ParamField path="audioRef" type="string"> Single reference audio (path or URL). Used for background music or voice reference when the provider supports audio inputs. </ParamField> <ParamField path="audioRefs" type="string[]">Multiple reference audios (up to 3).</ParamField> <ParamField path="audioRoles" type="string[]"> Optional per-position role hints parallel to the combined audio list. Canonical value: reference_audio. </ParamField>

Style controls

adaptive is a provider-specific sentinel: it is forwarded as-is to providers that declare adaptive in their capabilities (e.g. BytePlus Seedance uses it to auto-detect the ratio from the input image dimensions). Providers that do not declare it surface the value via details.ignoredOverrides in the tool result so the drop is visible.

Advanced

Reference inputs select the runtime mode:

• No reference media → generate
• Any image reference → imageToVideo
• Any video reference → videoToVideo
• Reference audio inputs do not change the resolved mode; they apply on top of whatever mode the image/video references select, and only work with providers that declare maxInputAudios.

Mixed image and video references are not a stable shared capability surface. Prefer one reference type per request.

Fallback and typed options

Some capability checks are applied at the fallback layer rather than the tool boundary, so a request that exceeds the primary provider's limits can still run on a capable fallback:

• Active candidate declaring no maxInputAudios (or 0) is skipped when the request contains audio references; next candidate is tried.
• Active candidate's maxDurationSeconds below the requested durationSeconds with no declared supportedDurationSeconds list → skipped.
• Request contains providerOptions and the active candidate explicitly declares a typed providerOptions schema → skipped if supplied keys are not in the schema or value types do not match. Providers without a declared schema receive options as-is (backward-compatible pass-through). A provider can opt out of all provider options by declaring an empty schema (capabilities.providerOptions: {}), which causes the same skip as a type mismatch.

The first skip reason in a request logs at warn so operators see when their primary provider was passed over; subsequent skips log at debug to keep long fallback chains quiet. If every candidate is skipped, the aggregated error includes the skip reason for each.

Actions

Model selection

OpenClaw resolves the model in this order:

1. model tool parameter — if the agent specifies one in the call.
2. videoGenerationModel.primary from config.
3. videoGenerationModel.fallbacks in order.
4. Auto-detection — providers that have valid auth, starting with the current default provider, then remaining providers in alphabetical order.

If a provider fails, the next candidate is tried automatically. If all candidates fail, the error includes details from each attempt.

Set agents.defaults.mediaGenerationAutoProviderFallback: false to use only the explicit model, primary, and fallbacks entries.

{
  agents: {
    defaults: {
      videoGenerationModel: {
        primary: "google/veo-3.1-fast-generate-preview",
        fallbacks: ["runway/gen4.5", "qwen/wan2.6-t2v"],
      },
    },
  },
}

Provider notes

Models: `seedance-1-0-pro-250528` (default),
`seedance-1-0-pro-t2v-250528`, `seedance-1-0-pro-fast-251015`,
`seedance-1-0-lite-t2v-250428`, `seedance-1-0-lite-i2v-250428`.

T2V models (`*-t2v-*`) do not accept image inputs; I2V models and
general `*-pro-*` models support a single reference image (first
frame). Pass the image positionally or set `role: "first_frame"`.
T2V model IDs are automatically switched to the corresponding I2V
variant when an image is provided.

Supported `providerOptions` keys: `seed` (number), `draft` (boolean —
forces 480p), `camera_fixed` (boolean).

Uses the unified `content[]` API. Supports at most 2 input images
(`first_frame` + `last_frame`). All inputs must be remote `https://`
URLs. Set `role: "first_frame"` / `"last_frame"` on each image, or
pass images positionally.

`aspectRatio: "adaptive"` auto-detects ratio from the input image.
`audio: true` maps to `generate_audio`. `providerOptions.seed`
(number) is forwarded.

Uses the unified `content[]` API. Supports up to 9 reference images,
3 reference videos, and 3 reference audios. All inputs must be remote
`https://` URLs. Set `role` on each asset — supported values:
`"first_frame"`, `"last_frame"`, `"reference_image"`,
`"reference_video"`, `"reference_audio"`.

`aspectRatio: "adaptive"` auto-detects ratio from the input image.
`audio: true` maps to `generate_audio`. `providerOptions.seed`
(number) is forwarded.

Provider capability modes

The shared video-generation contract supports mode-specific capabilities instead of only flat aggregate limits. New provider implementations should prefer explicit mode blocks:

capabilities: {
  generate: {
    maxVideos: 1,
    maxDurationSeconds: 10,
    supportsResolution: true,
  },
  imageToVideo: {
    enabled: true,
    maxVideos: 1,
    maxInputImages: 1,
    maxInputImagesByModel: { "provider/reference-to-video": 9 },
    maxDurationSeconds: 5,
  },
  videoToVideo: {
    enabled: true,
    maxVideos: 1,
    maxInputVideos: 1,
    maxDurationSeconds: 5,
  },
}

Flat aggregate fields such as maxInputImages and maxInputVideos are not enough to advertise transform-mode support. Providers should declare generate, imageToVideo, and videoToVideo explicitly so live tests, contract tests, and the shared video_generate tool can validate mode support deterministically.

When one model in a provider has wider reference-input support than the rest, use maxInputImagesByModel, maxInputVideosByModel, or maxInputAudiosByModel instead of raising the mode-wide limit.

Live tests

Opt-in live coverage for the shared bundled providers:

OPENCLAW_LIVE_TEST=1 pnpm test:live -- extensions/video-generation-providers.live.test.ts

Repo wrapper:

pnpm test:live:media video

This live file loads missing provider env vars from ~/.profile, prefers live/env API keys ahead of stored auth profiles by default, and runs a release-safe smoke by default:

• generate for every non-FAL provider in the sweep.
• One-second lobster prompt.
• Per-provider operation cap from OPENCLAW_LIVE_VIDEO_GENERATION_TIMEOUT_MS (180000 by default).

FAL is opt-in because provider-side queue latency can dominate release time:

pnpm test:live:media video --video-providers fal

Set OPENCLAW_LIVE_VIDEO_GENERATION_FULL_MODES=1 to also run declared transform modes the shared sweep can exercise safely with local media:

• imageToVideo when capabilities.imageToVideo.enabled.
• videoToVideo when capabilities.videoToVideo.enabled and the provider/model accepts buffer-backed local video input in the shared sweep.

Today the shared videoToVideo live lane covers runway only when you select runway/gen4_aleph.

Configuration

Set the default video-generation model in your OpenClaw config:

{
  agents: {
    defaults: {
      videoGenerationModel: {
        primary: "qwen/wan2.6-t2v",
        fallbacks: ["qwen/wan2.6-r2v-flash"],
      },
    },
  },
}

Or via the CLI:

openclaw config set agents.defaults.videoGenerationModel.primary "qwen/wan2.6-t2v"

Related

• Alibaba Model Studio
• Background tasks — task tracking for async video generation
• BytePlus
• ComfyUI
• Configuration reference
• fal
• Google (Gemini)
• MiniMax
• Models
• OpenAI
• Qwen
• Runway
• Together AI
• Tools overview
• Vydra
• xAI