Back to all models

MiniMax / Video

Hailuo.
Give the reaction room to land.

Hailuo is MiniMax’s video generation family. Its documented capabilities include character movement, stylised scenes and facial expression. That makes it a candidate for the close reactions and physical beats that carry much of a short episode’s emotion.

Inside the storyVideo
A close portrait of fictional character Anaya indoors, with rain in her hair.
AI Director’s shot briefShe recognises the handwriting.
01 / Performance

A small eye movement, then a held breath.

02 / Framing

A close-up that leaves room for the reaction.

03 / Edit

Hold long enough to read the thought before cutting.

Fictional story concept. Illustrative artwork, not output from the featured model.

Model capabilities

What Hailuo brings to the frame

Character action
Direct bodily movement and object motion as part of a generated video scene.
Facial expression
MiniMax describes support for subtle expression changes, which can be explored in reaction shots and close performance.
Stylised video
The family supports a range of visual treatments, including illustration and anime-oriented imagery. Test the treatment you intend to use.

Capabilities describe the model families, not a guarantee that every version or control is available in Tosheo. Production access depends on provider availability, references, rights and the approved scope.

02 / Hailuo

Where Hailuo fits in your story

Hailuo is a candidate for reaction coverage, physical gestures and stylised character moments. The AI Director can specify the emotional beat, the scale of the gesture and the cut point, using AI Skills to keep the request grounded in performance direction rather than a vague request for drama.

Fictional exampleShe recognises the handwriting.

Keep Anaya almost still. Let recognition arrive before she decides what to say.

Performance
A small eye movement, then a held breath.
Framing
A close-up that leaves room for the reaction.
Edit
Hold long enough to read the thought before cutting.

03 / Hailuo

Plan the Hailuo handoff

The AI Director prepares the video request from the scene plan, using AI Skills to specify the framing, performance or sound intention. You can review creative preparation or use Auto mode within your chosen scope. The model receives a bounded job, with the references and constraints that matter to that asset.

For an independent creator, this connects a creative decision to the next production step. For a publisher, it keeps recurring assets attached to the episode plan. Brand buyers can carry approved product references into the shot; production houses can retain the treatment and handoff requirements across the sequence.

04 / Hailuo

Where the automated line actually falls

QA can checkAgainst a reference
  • Is this the approved face and body
  • Is the costume right for this story day
  • Is the expression within the approved range
  • Does the dialogue match the approved script
  • Is the pronunciation right for this track
  • Is lip-sync within the treatment's tolerance
Only a person can judgeAgainst the moment
  • Does the hesitation read as hesitation
  • Is the restraint doing what the scene needs
  • Does the beat land at the right length
  • Is this the performance the episode was built for
  • Would a viewer believe it

This is the clearest illustration of why the human-control rule is a product decision and not a compliance gesture. Automation can propose, generate, check, route and assemble. Acceptance is a judgement, and quality in this system is defined as accepted output.

05 / Hailuo

Expression range is canon

What makes performance reviewable at all is that the character has an approved emotional vocabulary. An expression sheet guides the performance without changing who the character is - the same person in a different state, rather than a different person.

Without it, every performance note is a negotiation from scratch, and the season has no through-line because nothing recorded what this character's restraint looks like.

Fictional exampleThe hesitation

Kabir decides not to answer a simple question. A close-up has to carry it.

Inherited
Approved identity and expression range
Inherited
He knows who wrote the letter. She does not.
Checked
Identity, costume, dialogue, lip-sync tolerance
Judged
Whether the look reads as hesitation, by a person

A result can pass every automated check and still be rejected, because the check was never the point of the shot. That rejection is recorded against a named reviewer, and the repair is a shot rather than an episode.

06 / Hailuo

What acceptance would have to show

Provider documentation

Capabilities checked against official sources. Story applications are Tosheo’s editorial examples, not comparative benchmarks.

MiniMax: Hailuo character and motion capabilities

Questions before production

Can lip-sync quality be checked automatically?

Within a tolerance set by the treatment, yes, and it is checked per language track because the tolerance that works for one track may not survive another. What cannot be checked is whether the line was delivered well, which is a master-review decision.

Why does the treatment set a lip-sync tolerance at all?

Because it is a creative decision with production consequences. A treatment demanding tight sync narrows which providers can serve a shot and raises attempts per accepted shot, in every declared language. Deciding it once, at the treatment stage, is cheaper than discovering it per track.

What happens when a performance is rejected?

It is a targeted repair at shot level, with the same references and treatment, and the original is kept. The neighbouring accepted shots and the audio for the scene are untouched.

Do you generate performances of real actors?

No. Tosheo will not produce celebrity likeness or cloned voice without verified permission, and does not depict real people, actors or franchises in generated artwork.

Is expressive performance the deciding routing factor?

For close coverage in a character-driven genre, it weighs heavily. For an establishing shot it is irrelevant. That is the point of routing per unit rather than picking one provider for the whole production.

AI Director + AI Skills + Models

The model makes an asset. Your AI Director connects the story.

  1. 01

    Your direction

    Share the brief, format, episode plan and intended languages.

  2. 02

    AI Director

    Build the storyboard, references and shot-level direction with AI Skills.

  3. 03

    Model fit

    Match the asset to an available model and the approved production scope.

  4. 04

    Story-ready assets

    Check the result in context, repair what needs attention and assemble the edit.

Review creative preparation yourself or use Auto mode within your chosen scope. Cost authorisation, rights clearance and final delivery acceptance remain explicit decisions.

More from the model library

Keep building the story.

Tosheo / AI-native production

Start with the scene.
Let the direction connect it.

See how Tosheo brings the brief, AI Skills, generation and review into one production flow.