Act-One by Runway Review 2026: Is It Still Worth Using?

Quick Answer: Act-One by Runway is no longer the character-animation workflow new Runway users should build around in 2026. Act-One originally transferred facial expression, mouth movement and performance from a driving video to a character image or video using Gen-3 Alpha and Gen-3 Alpha Turbo. Runway retired those Gen-3 models in July 2026, and the current replacement is Act-Two, which expands performance capture with body movement, gestures and broader character support. If you're choosing a Runway performance-capture tool today, use Act-Two rather than Act-One.

Published: July 14, 2026
Last Updated: August 30, 2026

For our current evaluation of the full platform, read the Runway AI video generation review.

Act-One by Runway at a Glance

Act-One matters historically because it helped establish Runway's performance-driven character workflow, but it should no longer be treated as a standalone active product for new buyers.

What Was Act-One by Runway?

Act-One was Runway's performance-capture workflow for animating a character from a recorded human performance.

The basic process used two inputs:

1. a driving performance video, and
2. a character image or video.

The driving performance influenced elements such as:

• facial expression
• mouth movement
• pacing
• head movement
• character performance

while the character input determined the subject that appeared in the generated result.

This is different from traditional skeletal 3D motion capture.

Act-One was not primarily a system that generated rigged 3D characters or exported skeletal animation for game engines.

It was a generative video character-performance tool inside Runway.

That distinction is important when comparing it with tools such as DeepMotion.

How Act-One Worked

A typical Act-One workflow looked like this:

Performance video → Character reference → AI performance transfer → Generated character video

The driving video needed to show a clear performance.

Runway recommended:

• one visible face
• good lighting
• clear facial features
• forward-facing framing
• continuous visibility of the face
• clear mouth movements
• minimal occlusion
• no cuts during the driving performance.

For character images, Act-One worked best with realistic or semi-realistic human faces framed roughly shoulders-up.

That makes the old article's description of automatic full-body joint mapping and generalized animation rigging too broad.

Act-One Specifications

Runway's official Act-One documentation listed the following specifications.

These were credit costs, not separate Act-One monthly subscription prices.

Is Act-One Still Available in 2026?

For practical buying purposes, Act-One should now be treated as a legacy workflow.

Runway states:

• Gen-3 Alpha would no longer be available after July 8, 2026
• Gen-3 Alpha Turbo would no longer be available after July 30, 2026.

Act-One's documentation was built specifically around those models.

Meanwhile, Runway's current Apps directory lists:

Performance Capture with Act-Two

as its active character-performance workflow.

That means someone searching for Act-One today is better served by understanding what replaced it rather than being encouraged to subscribe specifically for Act-One.

What Replaced Act-One?

Runway's current replacement is Act-Two.

Act-Two follows the same broad concept but expands what can be transferred from the performance.

It supports:

• facial expression
• speech performance
• body movement
• hand gestures
• different character styles
• broader character angles
• character images
• character videos.

With a character image, Act-Two can transfer gestures and body movement from the driving performance.

It can also add environmental motion automatically to make a still character image feel more natural.

This is a meaningful upgrade over the more face-focused limitations of Act-One.

Act-One vs Act-Two

Act-Two currently generates at 5 credits per second and supports outputs up to 30 seconds.

Runway Pricing for Act-Two Users

Runway no longer sells Act-One as a separate $35+ SaaS product.

The correct way to understand pricing is through the broader Runway subscription.

Current monthly pricing is:

Annual billing currently reduces Standard, Pro and Max to approximately:

• Standard: $12/month
• Pro: $28/month
• Max: $76/month

Runway currently advertises a 20% annual discount.

For a detailed plan breakdown, read our Runway AI pricing guide.

Which plan gives access to Act-Two?

Act-Two is currently available on:

Standard or higher.

Runway does have a Free plan, but that should not be confused with free Act-Two access.

How Many Act-Two Videos Can Your Credits Generate?

Act-Two costs 5 credits per second.

If every included credit were used only on Act-Two, the theoretical capacity would be approximately:

Real workflows often use other models, iterations and editing tools, so actual monthly output will vary.

Runway's credit system is shared across image, video and audio generation.

What Act-One Did Well

1. Simple performance capture

Act-One lowered the barrier to character performance.

Instead of manually keyframing facial animation, creators could record a human performance and use it to drive another character.

2. Strong facial-expression transfer

Act-One's real strength was expression and mouth-performance transfer rather than full-body rigging.

For dialogue-driven character shots, that made it useful for:

• short narratives
• character tests
• social content
• concept videos
• animated dialogue.

3. Accessible workflow

Because the system lived inside Runway, users did not need a separate motion-capture pipeline or 3D-animation suite.

That made Act-One accessible to filmmakers and creators who wanted generative video rather than traditional animation engineering.

Act-One Limitations

1. Tight input requirements

Act-One worked best with clearly visible, forward-facing single subjects.

Occlusion, unusual movement and non-standard framing could reduce quality.

2. Limited body performance

Act-One should not be confused with modern full-performance transfer.

Its image-input guidance specifically favored relatively limited body/head motion.

Act-Two is considerably stronger for gestures and broader body performance.

3. Model dependency

Act-One depended on Gen-3 Alpha and Gen-3 Alpha Turbo.

Once Runway retired those models, the workflow ceased to be the sensible foundation for new projects.

4. It was not a 3D rigging tool

Users needing:

• FBX export
• skeletal animation
• game-engine rigs
• full mocap pipelines

should evaluate dedicated motion-capture or 3D-animation tools instead.

Who Was Act-One Best For?

At its peak, Act-One made the most sense for:

• filmmakers prototyping character scenes
• animators testing performance ideas
• short-form creators
• agencies producing stylized character content
• storytellers creating dialogue scenes
• artists experimenting with generative character performances.

It was less appropriate for:

• game-development motion capture
• complex full-body choreography
• long continuous scenes
• traditional 3D animation pipelines.

Who Should Use Act-Two Now?

In 2026, Act-Two is the better Runway choice when you need:

• expressive character animation
• driving-performance transfer
• facial performance
• body gestures
• hand movement
• stylized characters
• character-image animation.

Runway Academy currently teaches Act-Two specifically for expressive character performances, including facial expressions, gestures and voice workflows.

Act-One Alternatives in 2026