Integration · Developer tools

Connect Claude & ChatGPT to Runway

Ask your AI to generate a video, image or voice clip in Runway and it starts the task for you. It can make marketing images, product ad videos, UGC-style videos and localized ad images, run your published workflows and use your saved Model Routers. It also checks task status, reads the output links and tells you how many credits you have left.

Connect Runway free

Free plan, no credit card. Takes about a minute.

Try asking

“How many Runway credits do we have left?”
The AI reads your organization's usage tier, credit balance and usage.
“Make a 10-second product ad video from these three product photos.”
It starts a product ad video task with your images, ratio and duration, then checks the task until the video is ready and gives you the output link.
“Generate four marketing images of our coffee for Instagram, relaxed morning mood.”
It sends your brief to the marketing image tool with an output count of 4 and returns a task ID it follows with get_task.

What your AI can do

Video, images and audio

Generate video, images, speech and sound through your saved Model Routers, with a dry run that estimates cost first.

Product and ad content

Create marketing images, product campaign images, product ad videos, UGC-style videos, product swaps in video and localized ad images.

Tasks, workflows and credits

Check task status and outputs, run published workflows, list avatars and voices, and see your credit balance.

“Translate this ad image into Spanish and Japanese.”
It starts one localization task per language with your reference image and returns the task IDs.
“What would a 5-second video through our video router cost?”
It lists your Model Routers to find the config ID and runs generate_video with dry_run, which returns the estimated cost without generating anything.
“Run our product-shot workflow with this new photo.”
It lists your published workflows, reads the workflow's nodes and default inputs, runs it with your photo and checks the run until the outputs are ready.

65 tools for Runway

These tools are switched on when you connect. You can switch any of them off, or require your approval before it runs.

cancel_taskDeletesCancel a running or pending task by `task_id`; for finished tasks this permanently deletes the task and its outputs. Cannot be undone.
create_avatarChanges dataCreate a new avatar. Required: runway version, name, reference image, personality or voice. API summary: Create avatar.
create_avatar_videoChanges dataCreate a new avatar video. Required: runway version, model, avatar or speech. API summary: Generate avatar video from audio or text.
create_character_performanceChanges dataCreate a new character performance. Required: runway version. API summary: Control a character.
create_documentChanges dataCreate a new document. Required: runway version, name or content. API summary: Create document.
create_image_to_videoChanges dataCreate a new image to video. Required: runway version.
create_image_upscaleChanges dataCreate a new image upscale. Required: runway version.
create_multi_shot_videoChanges dataCreate a new multi shot video. Required: runway version. API summary: Create a multi-shot video.
create_realtime_sessionChanges dataCreate a new realtime session. Required: runway version. API summary: Create realtime session.
create_routerChanges dataCreate a new router. Required: runway version or slug. API summary: Create Model Router.
create_sound_effectChanges dataCreate a new sound effect. Required: runway version. API summary: Generate sound effects.
create_speech_to_speechChanges dataCreate a new speech to speech. Required: runway version.
create_text_to_imageChanges dataCreate a new text to image. Required: runway version. API summary: Text/Image to Image.
create_text_to_speechChanges dataCreate a new text to speech. Required: runway version.
create_text_to_videoChanges dataCreate a new text to video. Required: runway version.
create_usageChanges dataCreate a new usage. Required: runway version. API summary: Query credit usage.
create_video_to_hdrChanges dataCreate a new video to hdr. Required: runway version.
create_video_to_videoChanges dataCreate a new video to video. Required: runway version.
create_video_upscaleChanges dataCreate a new video upscale. Required: runway version.
create_voiceChanges dataCreate a new voice. Required: runway version, name or from. API summary: Create a voice.
create_voice_dubbingChanges dataCreate a new voice dubbing. Required: runway version.
create_voice_isolationChanges dataCreate a new voice isolation. Required: runway version.
delete_avatarDeletesPermanently delete a avatar by avatar id. Destructive action; may require approval. API summary: Delete avatar.
delete_documentDeletesPermanently delete a document by document id. Destructive action; may require approval. API summary: Delete document.
delete_messageDeletesPermanently delete a message by message id. Destructive action; may require approval. API summary: Delete conversation.
delete_realtime_sessionDeletesPermanently delete a realtime session by realtime session id. Destructive action; may require approval. API summary: Cancel realtime session.
delete_routerDeletesPermanently delete a router by router id. Destructive action; may require approval. API summary: Delete Model Router.
delete_voiceDeletesPermanently delete a voice by voice id. Destructive action; may require approval.
generate_audioChanges dataGenerate speech or audio through a saved Model Router: pass `config_id` and input.type ("speech" speaks promptText verbatim, "audio" treats it as a sound description) with input.promptText. dry_run true estimates cost only. Uses credits.
generate_imageChanges dataGenerate an image through a saved Model Router: pass `config_id` (router slug) and input.promptText, plus optional reference images, aspect ratio and more. dry_run true returns the routing decision and estimated cost without generating. Uses credits.
generate_marketing_imageChanges dataGenerate marketing stock images from a `prompt` brief (subject, audience, channel, mood). Set `version` (e.g. "2026-06"), optional output_count (1-4), quality and a reference_image (brand logo) uri. Uses credits; returns a task ID to check with get_task.
generate_product_ad_videoChanges dataCreate a product ad video from 1-10 `product_images` (uri each), with optional product_info, user_concept, style_images, ratio, duration (4-15 s) and audio. Set `version`. Uses credits; returns a task ID to check with get_task.
generate_product_campaign_imageChanges dataGenerate campaign images that keep a product intact: pass `image.uri` (product photo URL), a style `prompt` and `version`. Uses credits; returns a task ID to check with get_task.
generate_product_ugc_videoChanges dataCreate a UGC-style video of a person presenting a product: pass product_image.uri, character_image.uri and `version`, plus optional product_info, user_concept (tone or script), ratio, duration and audio. Uses credits; returns a task ID.
generate_videoChanges dataGenerate a video through a saved Model Router: pass `config_id` (router slug) and input with promptText and optional referenceImages, duration, resolution, aspectRatio and audio. dry_run true returns the estimated cost only. Uses credits.
get_audit_logGet one audit log by event id. API summary: Get a linked workspace audit log entry.
get_avatarGet one avatar by avatar id. API summary: Get avatar.
get_documentGet one document by document id. API summary: Get document.
get_messageGet one message by message id. API summary: Get conversation.
get_organizationGet the Runway organization behind the API key: usage tier, credit balance and usage. Use to check remaining credits before generating.
get_realtime_sessionGet one realtime session by realtime session id. API summary: Get realtime session.
get_routerGet one router by router id. API summary: Retrieve Model Router.
get_taskGet the status and outputs of a generation task by `task_id`. Poll this (no more than every few seconds) until the task succeeds, then read the output URLs.
get_voiceGet one voice by voice id. API summary: Get a voice.
get_workflowGet one published workflow by `workflow_id`, including its graph of nodes and their default inputs.
get_workflow_invocationGet the status and outputs of a workflow run by `workflow_invocation_id`.
list_audit_logsList audit logs. Optional filters: organization id, workspace ids, actions, actor emails, from or to. Results are paginated. API summary: List linked workspace audit logs.
list_avatar_usagesList avatar usages. Optional filters: start date or end date. API summary: Get avatar usage.
list_avatarsList the organization's avatars.
list_documentsList documents. Optional filters: sort or order. Results are paginated.
list_messagesList messages. Optional filters: avatar, start date or end date. Results are paginated. API summary: List conversations.
list_model_routersList the organization's saved Model Router configs. Their slugs are the config_id needed by generate_image, generate_video and generate_audio.
list_router_requestsList requests for a router. Results are paginated. API summary: List Model Router requests.
list_usagesList usages. Optional filters: from, to, organization id or workspace ids. Results are paginated. API summary: List linked workspace usage.
list_voicesList the custom voices in the organization.
list_workflowsList the organization's published Runway workflows.
localize_ad_imageChanges dataTranslate and localize an ad image into another language: pass reference_image.uri, target_language (e.g. "es", "ja") and `version`. Uses credits; returns a task ID.
run_workflowChanges dataRun a published workflow by `workflow_id`, optionally overriding inputs with node_outputs (node ID to output values). Uses credits; returns an invocation ID to check with get_workflow_invocation.
submit_voiceVisible to othersPreview a voice. Has effects visible outside this system.
swap_product_in_videoChanges dataReplace the product in a reference video: pass reference_video.uri, original_product_image.uri, new_product_images and `version`. Uses credits; returns a task ID to check with get_task.
update_avatarChanges dataUpdate an existing avatar identified by avatar id. Only the fields you pass are changed. API summary: Update avatar.
update_documentChanges dataUpdate an existing document identified by document id. Only the fields you pass are changed. API summary: Update document.
update_routerChanges dataUpdate an existing router identified by router id. Only the fields you pass are changed. API summary: Update Model Router.
update_voiceChanges dataUpdate an existing voice identified by voice id. Only the fields you pass are changed. API summary: Update a voice.
upload_a_fileChanges dataCreate a new document. Required: runway version, filename or type. API summary: Upload a file.

Set up in three steps

1
Pick the app
Create a free HeyMyApps account and choose Runway from the app list.
2
Paste your key
Paste your Runway API key (it starts with key_). Sign up in the Runway developer portal (dev.runwayml.com), create a project, open its API Keys tab and create a key. Copy it right away. Add credits to the project so generations can run.
3
Add the link to your AI
You get a personal MCP link. Add it to Claude, ChatGPT or Cursor:
1. Click your name, then Settings › Connectors › Add custom connector.
2. Paste your link as the Remote MCP server URL.
3. Switch it on from the + menu in a chat.

Questions

What can the AI do in Runway?

It can generate video, images and audio through your Model Routers, create marketing images, product campaign images, product ad videos, UGC-style videos and localized ad images, and swap a product in a video. It can run published workflows, check tasks and workflow runs, list avatars, voices, workflows and Model Routers, read your organization's credit balance and cancel tasks.

Does the AI see my Runway API key?

No. Your Runway API key is encrypted at rest and never shown to the AI. After you save it, it is not shown again, not even to you. The AI only sees the results of the tools it calls.

Can I control what the AI is allowed to do?

Yes. You choose which tools are switched on, so you can start with lookups only. Generation tools use credits, and cancelling a finished task permanently deletes the task and its outputs. These actions can require your approval before they run. Every tool call is logged.

What does it cost?

HeyMyApps has a free plan with no credit card required. Paid plans bill per completed task. You also need your own Runway developer account, and generations use the credits in your Runway project.

How do I get the finished video or image?

Runway generations run as tasks. The AI starts the task, gets a task ID back and checks it with get_task until it has succeeded. Then it reads the output URLs and gives you the links. Workflow runs work the same way with get_workflow_invocation.