All featuresAI Video Generation

Footage from a prompt. Rendered into your templates.

Clips from a prompt, an image or a talking portrait, rendered inside your templates from Studio, the API or an agent.

See plans and AI creditsRead the docs

On every paid plan · 3 AI credits per second of HD footage

AI clip layertext layersCTA layer5 srendered via API

The footage is one AI clip layer. Wordmark, headline, price and CTA are template layers on top, rendered as one MP4 by the Orshot API.

Prompt, image or portrait

A prompt for footage, an image to animate, or a photo plus a script to speak.

11 clip types, 37 looks

Pick what the clip does and how it looks. The prompt only describes the subject.

Generate per render

Any video layer accepts a prompt on the render. Repeats come from cache, free.

Studio, API and agents

One generator in Studio, one API endpoint, one tool for Claude, Cursor and ChatGPT.

Orshot Studio

Insert Video, then AI Generator.

Insert Video, AI Generator tab, a prompt, a clip type and a look. The clip lands on the canvas as a normal video layer.

  • 4 to 15 seconds. HD or Full HD
  • Aspect ratio follows the canvas
  • Regenerate for a fresh take
How AI video works in Studio
Talking photo

A portrait, a script, a presenter.

Turn on Make it speak, add a portrait and type the script. Orshot voices it and animates the face to match.

  • Script plus voice, or your own audio
  • Word timings come back, so captions sync
  • 60 to 180 seconds of speech depending on plan. Paid plans, with consent
Automate faceless videos with n8n
API

One endpoint, no mode to pick.

What you send decides what you get: a prompt, an image to animate, two frames to bridge, or a portrait plus a script.

  • URLs, data URIs, brand:image:<id> or gen://<hash> as inputs
  • Estimate first with /v1/ai/video/estimate, free
  • Poll the job, pass sync, or set a webhook
API reference
Generate a clip
curl -X POST https://api.orshot.com/v1/ai/video \
  -H "Authorization: Bearer $ORSHOT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "black earbuds case on stone, slow orbit",
    "preset": "product-hero",
    "look": "clean-studio",
    "aspect": "16:9",
    "duration": 5
  }'
Make a photo speak
{
  "imageRef": "brand:image:34523",
  "script": "Three reasons this launch matters.",
  "voice": "XrExE9yKIg1WjnnlVkGX",
  "consent": true,
  "preset": "ugc-selfie-talk"
}
Render-time generation

Add .prompt to a video layer. Done.

Pass the prompt on the render itself. The clip is generated, cached and composited into the template in one call.

  • Same prompt, same clip: it generates once, then caches
  • Reuse any clip with gen://<hash>, no credits spent
  • Add .prompt to text and image layers too
AI content at render time
curl -X POST https://api.orshot.com/v1/studio/render \
  -H "Authorization: Bearer $ORSHOT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "templateId": 18917,
    "modifications": {
      "headline": "Meet the new case",
      "hero_clip.prompt": "black earbuds case, slow orbit",
      "hero_clip.preset": "product-hero",
      "hero_clip.look": "clean-studio"
    },
    "response": { "format": "mp4", "mode": "async" }
  }'
API use cases

The requests people actually send.

Complete requests for the most common jobs, from a first text to video call to reusing a cached clip inside a render.

Text to video

Footage from a prompt. Returns a job; poll it, then use the clip URL anywhere.

Docs
curl -X POST https://api.orshot.com/v1/ai/video \
  -H "Authorization: Bearer $ORSHOT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "prompt": "a ceramic coffee cup on a wooden table by a window, morning light",
  "duration": 5,
  "preset": "product-hero",
  "look": "clean-studio",
  "aspect": "9:16"
}'

Same request shape from n8n, Zapier, Make and the agent tools.

Image to video

Animate a still. imageRef is the first frame; references guide style or product.

Docs
curl -X POST https://api.orshot.com/v1/ai/video \
  -H "Authorization: Bearer $ORSHOT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "prompt": "slow push-in, steam rising",
  "imageRef": "brand:image:123",
  "references": [
    "brand:image:124"
  ],
  "duration": 5,
  "preset": "cinematic-scene"
}'

Same request shape from n8n, Zapier, Make and the agent tools.

Transition between two frames

Give a first and a last frame and the model fills the motion between them.

Docs
curl -X POST https://api.orshot.com/v1/ai/video \
  -H "Authorization: Bearer $ORSHOT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "prompt": "the morning street becomes the same street at night",
  "imageRef": "brand:image:123",
  "lastImageRef": "brand:image:125",
  "duration": 5,
  "preset": "transition"
}'

Same request shape from n8n, Zapier, Make and the agent tools.

Talking photo

A portrait speaks a script. The script runs through text to speech first, then the face is animated.

Docs
curl -X POST https://api.orshot.com/v1/ai/video \
  -H "Authorization: Bearer $ORSHOT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "imageRef": "brand:image:123",
  "script": "Three reasons this launch matters.",
  "voice": "XrExE9yKIg1WjnnlVkGX",
  "consent": true,
  "preset": "ugc-selfie-talk"
}'

Same request shape from n8n, Zapier, Make and the agent tools.

Generate inside a template render

Dotted keys on any video layer of your template. The clip is generated, cached and rendered into the design in one call.

Docs
curl -X POST https://api.orshot.com/v1/studio/render \
  -H "Authorization: Bearer $ORSHOT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "templateId": 1234,
  "modifications": {
    "headline": "Meet the new case",
    "hero_clip.prompt": "black earbuds case on stone, slow orbit",
    "hero_clip.preset": "product-hero",
    "hero_clip.look": "clean-studio"
  },
  "response": {
    "format": "mp4",
    "mode": "async"
  }
}'

Same request shape from n8n, Zapier, Make and the agent tools.

Reuse a generated clip for free

Pass the gen:// reference of an earlier clip as the layer value. No credits are spent.

Docs
curl -X POST https://api.orshot.com/v1/studio/render \
  -H "Authorization: Bearer $ORSHOT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "templateId": 1234,
  "modifications": {
    "headline": "Back in stock",
    "hero_clip": "gen://7e8977b9911eaaefe2e23a25dec53cbf45136178"
  },
  "response": {
    "format": "mp4",
    "mode": "async"
  }
}'

Same request shape from n8n, Zapier, Make and the agent tools.

Estimate credits before generating

Same body as a generation. Returns the credits it would cost without spending any.

Docs
curl -X POST https://api.orshot.com/v1/ai/video/estimate \
  -H "Authorization: Bearer $ORSHOT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "prompt": "a sneaker drops onto wet asphalt at night",
  "duration": 10,
  "fullHd": true
}'

Same request shape from n8n, Zapier, Make and the agent tools.

List clip types and looks

Every role and look preset with preview clips, plus the credit rate card.

Docs
curl https://api.orshot.com/v1/ai/presets \
  -H "Authorization: Bearer $ORSHOT_API_KEY"

Same request shape from n8n, Zapier, Make and the agent tools.

Poll a generation job

Generation returns a job. Poll it until finished is true, or pass sync or webhook_url on the request instead.

Docs
curl https://api.orshot.com/v1/studio/render-jobs/89 \
  -H "Authorization: Bearer $ORSHOT_API_KEY"

Same request shape from n8n, Zapier, Make and the agent tools.

Orshot for agents

Ask Claude for a clip. Get it in the template.

Connect Orshot for agents and Claude, Cursor or ChatGPT list the presets, generate the clip, place it and render.

  • Tools: orshot_get_ai_video_presets, orshot_generate_ai_video
  • The agent reads the rate card before it spends
  • OAuth sign-in, no keys to wire
Explore Orshot for agents
Agent session
Completed
You

“Generate a 5 second product hero clip of the matte black earbuds case, clean studio look, and put it behind the headline in the launch template.”

Agent
orshot_get_ai_video_presets
orshot_generate_ai_video
orshot_patch_template_elements
AI generated clip of a matte black earbuds case on grey stone
1 / 1
Template #18917 · Launch teaserproduct-hero · clean-studio · 16:9 · 15 AI credits
Looks

37 looks, one word each.

A look sets the style, so the prompt only describes subject and motion. Every look ships with a preview clip; here are 18 of them.

Cinematic look preview: a coffee cup rendered in the Cinematic style
Cinematic
Clean studio look preview: a coffee cup rendered in the Clean studio style
Clean studio
Archival B&W look preview: a coffee cup rendered in the Archival B&W style
Archival B&W
Macro look preview: a coffee cup rendered in the Macro style
Macro
Anime look preview: a coffee cup rendered in the Anime style
Anime
Claymation look preview: a coffee cup rendered in the Claymation style
Claymation
Paper cut-out look preview: a coffee cup rendered in the Paper cut-out style
Paper cut-out
Neon look preview: a coffee cup rendered in the Neon style
Neon
Minimal 3D look preview: a coffee cup rendered in the Minimal 3D style
Minimal 3D
Watercolour look preview: a coffee cup rendered in the Watercolour style
Watercolour
Retro VHS look preview: a coffee cup rendered in the Retro VHS style
Retro VHS
Storybook look preview: a coffee cup rendered in the Storybook style
Storybook
Pixel art look preview: a coffee cup rendered in the Pixel art style
Pixel art
Pop art look preview: a coffee cup rendered in the Pop art style
Pop art
Cyberpunk look preview: a coffee cup rendered in the Cyberpunk style
Cyberpunk
Product commercial look preview: a coffee cup rendered in the Product commercial style
Product commercial
Food look preview: a coffee cup rendered in the Food style
Food
Travel look preview: a coffee cup rendered in the Travel style
Travel
Brand Library and logs

Every clip is kept, every credit is listed.

New clips are saved to the Brand Library with their tags. Every generation is logged with its source and credits.

  • Filter the library by the ai tag or clip type
  • Logs show the source: Studio, API, render or agent
  • A daily cap per workspace stops a runaway automation
Brand Assets Library
AI credits

Billed per second, on delivery.

AI video, images, voice and the design agent share one credit pool. Every plan includes credits. Top up 200 for $11.

  • Paid plans include 300 to 3,600 AI credits a month. AI video is not on the free plan
  • Charged on delivery. Failed and cached generations are free
Compare plans
AI video, HD (up to 768p)
3 credits / second
AI video, Full HD (1080p)
6 credits / second
Talking photo
3 credits / second + 1 or 2 per 10 s of speech
Voiceover or ambient audio
1 credit per started 10 s
Premium voices
2 credits per started 10 s
AI image
2 credits
Background removal
1 credit

What 50 free credits buy

  • Three 5 second HD clips
  • Or one 6 second clip plus 16 AI images
  • Or 25 AI images

A 5 second HD clip is 15 credits. Free clips run up to 6 seconds; voiceovers and talking presenters start on paid plans. Top up 200 credits for $11 on any paid plan. Charged only when a clip is delivered.

AI audio generation

Voiceovers and ambient sound for the same templates, with word-synced captions.

Video generation API

Render MP4, WebM, MOV or GIF from a template with trim, audio, subtitles and transitions.

Credits and usage

What a render credit and an AI credit cover, and how each output is billed.

FAQ

Questions about AI video generation.

What people ask before they generate their first clip through Orshot.

What is an AI video generation API?

An API that turns a text prompt, an image or a portrait into a short video clip. Orshot exposes one endpoint, POST /v1/ai/video, and adds what most AI video APIs leave out: the clip is saved to your Brand Library and can be rendered inside a template with text, logo and CTA layers on top.

How does AI video generation work in Orshot?

You send what you have. A prompt returns footage, a prompt plus an image animates that image, two images produce a transition, and a portrait plus a script produces a talking presenter. Pick a clip type and a look so the prompt stays short. The clip comes back in 15 to 45 seconds for a 5 second HD render and lands in the Brand Library, Studio, the API and your AI agent alike.

Can I generate an AI video from text only?

Yes. A prompt on its own returns footage from text. Add an image when you want a specific product or scene animated, and add a script and a portrait when you want a person to speak.

Is there a free AI video generator API?

AI video is included on every paid plan, from Launch up. Plans include 300 to 3,600 AI credits a month, and a 5 second HD clip is 15 credits, so the smallest plan covers about 20 clips before a top up. The free plan keeps 50 AI credits for AI images and the design agent, but does not generate video.

How much does AI video generation cost?

3 AI credits per second in HD and 6 per second in Full HD, charged only when a clip is delivered. A 5 second HD clip is 15 credits. Failed generations cost nothing and identical inputs are served from cache for free. You can top up 200 credits for $11 on any paid plan.

How long can an AI generated clip be?

Footage clips run up to 15 seconds; the shortest clip the models deliver is about 4 seconds. Talking presenter clips follow the length of the script, from 60 seconds on Launch up to 180 seconds on Scale.

Which AI video models does Orshot use?

Orshot routes each request to the model that fits the capability and resolution tier, with a fallback when a provider is down. You call one endpoint and pick HD or Full HD; you never manage model keys, and your requests do not change when the models behind them do.

Can AI agents like Claude or ChatGPT generate videos with Orshot?

Yes. Connect Orshot for agents and Claude, Cursor, ChatGPT or Codex can list the presets, generate a clip, place it in a template and render the result through the same tools the Studio uses. Sign-in is OAuth, so there are no API keys to wire.

How is this different from an AI video generator like Runway or Canva?

Those tools generate a clip you then edit by hand. Orshot treats the clip as one layer of a template, so headline, price, logo and captions render on top automatically, and the whole thing can be produced per customer, per product or per campaign through the API or a workflow.
Ready to generate?

Prompt in. Rendered video out.

AI credits are included on every plan. Pick the one that matches your render volume.

See plansTalk to us