AI videos where your actors never break character

Create an actor from a photo, in the constructor or with a prompt — the face and look persist in every scene. Shoot films of any length scene by scene — or recast any video with your own actors.

Real people only with confirmed consent · sound and 2K upscaling · API & MCP for agents

DEMO · made with Actoria

Recast: your face in any clip

Upload a clip — the movement, camera, timing and the original lines stay; the face and the look are yours.

Original clip
Recast — the same scene, a new face

Made in Actoria: a film scene as the source, a verified user as the actor, the original soundtrack kept.

Original clip
Recast — the same scene, a new face

Made in Actoria: a dance scene as the source, a verified user as the actor, every move and the soundtrack kept.

Your actors, in order of appearance
Up to 4 actors from your library are mapped to the people in the source, first to last. Add a prop and a location to change what's in frame and where it happens.
Up to 30 seconds per recast
MP4 or MOV, 4–60 seconds; the first 30 seconds are recast. The result is as long as the source and can keep its original soundtrack.
Faces stay private
Faces in your upload are blurred before the video leaves our server — the model sees the motion, not the faces.
Try RecastAvailable on any paid plan
01
Actor
Verify your own face, invite friends by link — or build a fictional actor in the constructor or with a prompt.
02
Scene
Pick or describe the setting and props: interior or street, time of day, weather, mood — and the objects in frame, from boxing gloves to a sports car.
03
Video
Describe the action, set the camera — and watch the frame come alive. Carry on from the last frame as many times as you like.

Shoot the whole film — no time limit, same cast in every scene

Add scenes for as long as the story needs — there is no cap on their number. Continue picks up the next scene from the last frame of the previous one, with the same actors, props, location and camera. Finish assembles every scene into one film with sound.

Same actors, props and locations
Your cast lives in a library, not in a prompt: the same faces, objects and places appear in the first scene and in the last.
No time limit
Scenes run up to 30 seconds each, and you can add as many as you like — a trailer, an episode or a whole film.
One film, with sound
Finish stitches all scenes into a single video with its soundtrack — ready to download and share.
Made in Actoria
Three scenes generated in Actoria with one actress and assembled with Finish into one video with sound.

Why creators pick Actoria

You and your friends star in it
A selfie liveness check verifies your face; friends confirm consent on their own screen — and everyone lands in your actor library.
Consistent cast, props and locations
Faces, locations and props persist from scene to scene — no random look-alikes.
Video from the first frame
Photo first: perfect the opening frame, then turn it into video — no slot-machine generations.
Films of any length
No cap on the number of scenes: Continue starts the next one from the last frame with the same cast, Finish assembles them all into one film with sound.
Recast any scene
Upload a video and get it back with your actors, props and location — same motion, camera, timing and dialogue. Up to 30 seconds.
Cinematic quality
HD video with smooth motion and camera control, upscaling to 2K.
API & MCP

API & MCP: run Actoria from your own tools

Connect Claude Code, Claude.ai, Cursor or the Claude API to your account over the Model Context Protocol. 28 tools cover the whole pipeline — actors, first frames, videos, script-to-film and audio — with the same limits and prices as on the site.

28 toolsClaude Code · Claude.ai · Cursor · Claude APIKeys scoped read / generate / upload
claude mcp add --transport http actoria https://actoria.ai/mcp \
  --header "Authorization: Bearer act_live_…"

From the blog

All articles →

An AI video generator built for stories

Actoria is an online AI video generator with consistent characters: create an actor from a photo, in the constructor or with a prompt, and they keep the same face and look in every scene. That's the piece ordinary text-to-video tools miss — where every new clip stars a new stranger.

The workflow is built around the frame: image-to-video brings to life the exact first frame you approved, and any clip can be extended from its last frame — short fragments grow into long stories: trailers, episodes, entire AI films.

There is no cap on the number of scenes in a project: Continue starts the next scene from the last frame with the same actors, props and location, and Finish assembles every scene into one film with sound — the only ceiling is your plan's video storage. Recast adds video-to-video: upload a clip, and the result repeats its motion, camera, timing and dialogue with your cast, up to 30 seconds at a time. All of it is also available over API & MCP — from Claude Code, Claude.ai, Cursor or the Claude API.

Real people appear on screen only with consent: you verify your own look with a selfie liveness check and invite friends by link. Finished videos come with sound and a 2K upscale.

Frequently asked questions

How do I keep the same character in every scene of an AI video?

Create an actor once — from a photo or in the builder. Actoria uses it as a reference for every generation, so the face and look stay the same from scene to scene.

How long can an AI film be in Actoria?

As long as you need — there is no cap on the number of scenes in a project. Each scene is up to 30 seconds; Continue starts the next one from the last frame with the same cast, and Finish assembles them all into one film with sound. The only ceiling is your plan's video storage: 100 MB on Basic, 1 GB on Premium.

How do I turn an image into a video?

Compose the first frame from an actor and a location (or upload your own), describe the action and the camera — Actoria brings the frame to life as video.

Can I create an AI actor from my own photo?

Yes — upload a photo and the actor keeps that look in every shot. You can only upload photos of yourself: identity is confirmed with a liveness check.

Do AI videos come with sound?

Yes — pick the sound-enabled model and the clip is generated together with its audio track.

What resolution are the videos?

Scenes generate in HD, and finals up to 45 seconds can be upscaled to 2K.

Can I make AI videos together with friends?

Yes — send your friends an invite link: each confirms their own appearance and consent, and then you can compose group frames and bring them to life.

What is Recast?

Recast is video-to-video: upload a clip or pick one of your finished videos, and Actoria shoots it again with your actors, props and location. The motion, camera, timing and dialogue come from the source; the first 30 seconds are recast, and the result is as long as the source. Available on any paid plan.

Can I use my own footage in Recast — and what happens to the faces in it?

Yes — MP4 or MOV, 4 to 60 seconds, as long as you have the rights to it. Faces in the upload are blurred on our server before the video goes to the model; the original never leaves our server.

Can I use Actoria from Claude, Cursor or my own code?

Yes — Actoria has an MCP server at actoria.ai/mcp with 28 tools: actors, frames, videos, script-to-film and audio. Create an API key in Settings and connect Claude Code, Claude.ai, Cursor or the Claude API; the docs are at actoria.ai/mcp/docs.