Runway ML

Overview

Use the runway/ model prefix for image and video operations. The gateway translates requests to Runway’s task API. Image generation and editing wait for the task result within the configured provider timeout; video generation returns a task ID for subsequent status and content requests.

Supported Operations

Operation	Supported	DeepIntShield endpoint
List Models	✅	GET /v1/models (bundled snapshot and configured IDs)
Image Generation	✅	POST /v1/images/generations
Image Editing	✅	POST /v1/images/edits (reference images; no mask)
Video Generation	✅	POST /v1/videos
Video Retrieve	✅	GET /v1/videos/{id}
Video Download	✅	GET /v1/videos/{id}/content
Video Delete	✅	DELETE /v1/videos/{id} (cancel)
Video List	❌	-
Video Remix	❌	-

Chat, Responses, image variations, and image streaming are not implemented by the Runway adapter. The bundled model metadata distinguishes image and video models and supplies known sizes and required reference inputs. Unknown IDs remain usable without guessed capabilities or defaults. The model snapshot does not verify your Runway account’s access.

Image generation and editing

curl https://app.deepintshield.com/v1/images/generations \
  -H "Authorization: Bearer $DEEPINTSHIELD_VIRTUAL_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "runway/gen4_image",
    "prompt": "A watercolor landscape at sunrise",
    "size": "1024x1024",
    "n": 1,
    "response_format": "b64_json"
  }'

Parameter	Behavior
prompt	Required text instruction.
size	Translated from WxH to Runway’s ratio field; supported values depend on the model.
input_images	Reference image URLs or data URLs for generation. Required for gen4_image_turbo.
reference_images	Reference objects preserving native URI/tag fields.
n	Must be positive. gen4_image and gen4_image_turbo accept one image per request.
response_format	b64_json downloads and returns encoded image bytes; otherwise results contain image URLs.

For edits, send multipart form data to /v1/images/edits with model, prompt, and image. Uploaded images become Runway references. The adapter rejects mask; it cannot apply a masked inpainting request. The Playground hides this unsupported control.

Both image operations poll Runway internally and return the final data[] result. A failed or cancelled task returns an error. If a valid image task exceeds the provider timeout, increase that provider’s request timeout and the caller’s timeout as needed.

---

Video generation

Generate (POST /v1/videos)

Request Parameters

Parameter	Type	Required	Notes
model	string	✅	Provider-qualified Runway model, such as runway/gen4.5
prompt	string	✅	Text description of the video
input_reference	string	Model-specific	Image URL or data URL for image-to-video; required for gen4_turbo.
seconds	string	Model-specific	Whole-number duration, such as "5". Defaults and allowed values are model-specific.
size	string	Model-specific	Resolution or ratio supported by the model, such as 1280x720.
seed	int	❌	Sent when supplied; support depends on the model.
audio	bool	❌	Enable audio generation on a model that supports it.
video_uri	string	❌	Source video URL for a supported video-to-video model.

Extra Params

Key	Type	Notes
references	array	Video reference objects [{"uri": "...", "tag": "..."}] for video-to-video
content_moderation	object	Content moderation config
reference_images	array	Reference image objects for style/asset guidance

Generation Modes (auto-detected from inputs)

• Text-to-video: prompt only
• Image-to-video: prompt + input_reference
• Video-to-video: prompt + video_uri, when the selected model supports it

Response: DeepIntShieldVideoGenerationResponse with id, status, videos[]

DeepIntShield statuses (normalized): queued → in_progress → completed / failed

These values are the normalized view returned by DeepIntShield’s API. Runway’s native statuses are: PENDING, THROTTLED, RUNNING, SUCCEEDED, FAILED, CANCELLED.

Retrieve / Download / Delete

Operation	Endpoint	Notes
Get status	GET /v1/videos/{id}	Poll until status: completed
Download content	GET /v1/videos/{id}/content	Returns raw video bytes (MP4)
Cancel/Delete	DELETE /v1/videos/{id}	Cancels the running task

---

Configuration

Add Runway as a provider from the dashboard:

1. Go to https://app.deepintshield.com and open Providers in the sidebar.
2. Click Add New Provider and select Runway as the provider type.
3. Add a key (for example, named runway-key-1) and paste your Runway API key.
4. Save the configuration.

Once saved, route requests to Runway through the gateway using your virtual key.

See Provider Configuration for full setup options.

---

Reference Links

• Runway ML API Documentation
• Runway ML Models