What Is CapCut AI? Features, Pricing, and Limits Explained

CapCut AI is the set of generative AI tools built into ByteDance's CapCut video editor, including script-to-video generation, AI avatars, voice cloning, auto-captions, and AI image generation. It is not a standalone product; it is a feature layer inside a free template-based editor, and its outputs are billed through a monthly AI credit allowance.

TL;DR

CapCut AI turns a text prompt or script into a multi-scene narrated video, stitching stock or AI-generated clips with voiceover, captions, and music.
It runs inside the free CapCut app; the free tier caps most AI features and limits exports to 1080p.
CapCut Pro costs $19.99/month (or $179.99/year) as of early 2026, up from the roughly $8-10/month plan long-time users remember.
AI credits are metered separately: avatar generation runs about 20-40 credits per video, voice cloning about 10-25 credits per minute, and script-to-video generation about 30-80 credits depending on length.
CapCut's June 2025 terms of service update grants ByteDance a broad, perpetual license to content uploaded to its cloud, including use for AI training, with no opt-out.
CapCut is built for editing and templating; it does not build a full video from a plain-language brief the way an end-to-end AI video agent does.

What Is CapCut AI?

CapCut AI is not a separate app. It is a bundle of generative features layered onto the existing CapCut editor, which itself started as a free mobile and desktop video editing tool built around templates, transitions, and stock effects.

What CapCut AI includes, as of 2026:

AI Script-to-Video / Long-Video Generator: takes a written script or topic prompt and assembles a multi-scene video using stock footage, AI voiceover, and auto-generated captions.
AI Avatars: generates a talking digital presenter from a photo or preset character, paired with a script and synthetic voice.
Voice Cloning: clones a voice from a 60-second sample and lets you script new voiceover in that cloned voice, in multiple languages.
AI Auto-Captions: generates subtitles in 20+ languages directly from audio.
AI Image Generation: Seedream 5.0 for text-to-image, with an early rollout of ByteDance's Dreamina Seedance 2.0 for AI video clips inside the app.
AI Auto-Edit and Auto-Cut: automatically trims raw footage into a paced sequence.
Background Removal: removes video backgrounds without a green screen, on short clips in the free tier.

What CapCut AI is not:

It is not an autonomous video agent that plans shots, routes generation across multiple AI models, and hands you a finished, scored video from one instruction.
It is not primarily a generation tool; its core product is still manual timeline editing, and AI features sit on top as add-ons.
It does not build videos from a URL, a product page, or an audio track as a starting input, the way some newer AI video agents do.

Where CapCut AI Came From

CapCut launched in 2020 as ByteDance's free spinoff of its Chinese app Jianying, aimed at TikTok creators who needed fast, template-driven mobile editing.
Through 2023-2024, CapCut added early AI features: auto-captions, background removal, and basic text-to-speech, positioning them as conveniences bolted onto the editor rather than a generation engine.
In 2025-2026, CapCut expanded aggressively into generative AI, adding script-to-video, AI avatars, and voice cloning as ByteDance pushed to compete with dedicated AI video tools.
In March 2026, ByteDance rolled out Dreamina Seedance 2.0 video generation inside CapCut, after briefly pausing the launch in mid-March over Hollywood copyright complaints, then resuming with added content safeguards.
On June 12, 2025, CapCut updated its terms of service to grant itself a non-exclusive, worldwide, perpetual, irrevocable, sublicensable license to content uploaded to its cloud, a change that drew significant creator and legal pushback.

How CapCut AI Works in Practice

You write or paste a script, or type a topic prompt, into the Script-to-Video tool.
CapCut's AI breaks the script into scenes and matches each scene to stock footage, template clips, or (where enabled) AI-generated video clips.
The tool adds AI voiceover (stock voice or your cloned voice), auto-generated captions, background music, and transitions.
You review the assembled draft on CapCut's timeline and manually adjust pacing, swap clips, or re-time cuts, since the AI output is a starting draft, not a final edit.
You export at up to 1080p on the free tier, or up to 4K/HDR on the Pro plan.
Each AI-heavy step (avatar generation, voice cloning, long-video generation) consumes AI credits from your monthly allowance rather than being unlimited.

For a full breakdown of the interface, see this CapCut tutorial, and for editing shortcuts once you're inside the app, see CapCut tips and tricks.

How CapCut AI Differs From Manual Editing (and From an AI Video Agent)

	Manual CapCut editing	CapCut AI features	An AI video agent (e.g., Pexo)
Starting point	Your own footage	A script or prompt, editable after generation	Plain-language description, script, URL, images, or audio
Output	Fully manual timeline	AI-assembled draft you still edit	Finished, edited, scored video
Model access	None (editor only)	ByteDance's own models (Seedream 5.0, early Seedance 2.0)	Auto-routes each shot across 10+ models including Seedance 2.0 and Kling AI
Audio	Manual music/voice add	AI voiceover + captions	Three-layer soundtrack: voiceover, music, and Foley sound effects
Skill required	Editing skills	Light prompting + manual cleanup	None; describe the outcome
Cost model	Free	Free tier capped; Pro $19.99/mo + metered AI credits	Usage-based, no separate credit system for editing skills
Content license	N/A for local edits	Broad ByteDance content/AI-training license on cloud uploads	Standard usage terms

Who CapCut AI Is For

TikTok and Reels creators who already live inside CapCut's mobile app and want auto-captions or quick background removal without leaving it.
Solo creators on a tight budget who can tolerate the free tier's 1080p cap and limited AI credits.
People repurposing existing footage, since CapCut's core strength is still template-based editing of clips you already shot.
Casual script-to-video experiments, where a rough AI-assembled first draft is acceptable and you plan to manually finish the edit yourself.

CapCut AI is a weaker fit for anyone who wants a single instruction to produce a finished, polished video with no manual timeline work, or who needs full audio design (voiceover, music, and sound effects together) rather than a bare AI voiceover track.

Pricing

Free tier: basic AI auto-captions, AI auto-cut, background removal on short clips, 1080p export.
Standard (renamed from the old Pro tier in early 2026): about $10/month.
Pro: $19.99/month, or $179.99/year paid annually. Includes 4K/HDR export, the full AI toolkit (avatars, voice cloning, text-to-video, camera tracking, vocal isolation), 1TB cloud storage, and 1,200 AI points per month.
AI credit costs: roughly 20-40 credits per AI avatar video, 10-25 credits per minute of cloned voice, and 30-80 credits per script-to-video generation depending on length and scene count. Heavy AI use can burn through the monthly allowance well before the month ends, requiring extra credit purchases.

See a full breakdown of what's actually behind the paywall in CapCut free vs. paid.

Limitations

AI output needs manual cleanup. Script-to-video and avatar generation produce a rough draft, not a finished piece; you still edit the timeline.
Credits run out fast. The metered AI credit system means heavier AI use (long videos, extended voice cloning) can exceed the monthly allowance and require top-ups.
No model choice. CapCut AI runs on ByteDance's own models (Seedream 5.0, an early rollout of Seedance 2.0); you cannot route a shot to a different AI video model.
Audio is a single voiceover track, not a full soundtrack with separate music and sound-effect layers.
Terms of service risk. CapCut's June 2025 terms update grants ByteDance a non-exclusive, worldwide, perpetual, irrevocable, sublicensable license to any content uploaded to its cloud, including use for AI training and no listed opt-out. This applies once content touches CapCut's cloud, not to material edited and exported without ever being uploaded. Creators and agencies handling client footage should read this clause carefully before using cloud sync.
No non-editing inputs. CapCut AI cannot start from a URL or a standalone audio track as a video's basis.

Alternatives to CapCut AI

If CapCut AI's rough-draft-plus-manual-editing model, its credit limits, or its content terms are a dealbreaker, worth comparing:

Full editors with different AI feature sets and pricing structures: see best CapCut alternatives.
An end-to-end AI video agent for people who want to skip manual editing entirely: describe a video in plain language, or hand over a script, images, a URL, or an audio track, and get back a finished, edited, scored video with auto model selection across 10+ models (including Seedance 2.0 and Kling AI) and a three-layer soundtrack (voiceover, music, and Foley sound effects). This is a genuinely different category from a template editor with AI add-ons; see vibe-creating vs. CapCut for how the two approaches actually compare.

Frequently Asked Questions

Is CapCut AI free? Yes, in a limited form. The free tier includes basic auto-captions, auto-cut, short-clip background removal, and 1080p export. Heavier AI features like avatars, voice cloning, and long-video generation require a paid plan and consume metered AI credits.

How much does CapCut AI cost? CapCut Pro is $19.99/month or $179.99/year as of early 2026, including 1,200 AI points per month. A cheaper Standard tier (around $10/month) sits below it without the full AI toolkit.

Does CapCut AI generate a full video from just a script? It generates a draft: scenes matched to stock or AI clips, plus AI voiceover and captions. You typically still adjust pacing and swap clips manually before it's finished.

Does CapCut own the videos I make? CapCut doesn't claim ownership of locally edited, non-uploaded work. But its June 2025 terms of service grant ByteDance a broad, perpetual, sublicensable license to anything uploaded to its cloud, including for AI training, with no stated opt-out.

What AI video model powers CapCut? CapCut AI runs on ByteDance's own models: Seedream 5.0 for images, and an early rollout of Dreamina Seedance 2.0 for AI video clips, added in March 2026 after a brief pause over copyright concerns.

Is CapCut AI the same as an AI video generator like Pexo? No. CapCut is a free template editor with generative AI features added on top; you still assemble and edit the output yourself. An AI video agent like Pexo takes a plain-language description (or a script, images, URL, or audio) and returns a finished, edited video without manual timeline work.

Do I need editing skills to use CapCut AI? Basic ones. AI features generate a starting draft, but getting a polished final result still requires editing skills to trim, re-time, and refine the timeline.

What happens if I run out of AI credits? You'll need to wait for the next monthly allowance or purchase additional credits; credit-metered features like avatars and voice cloning stop working once your balance is used up.

Conclusion

CapCut AI is a genuinely useful set of generative add-ons bolted onto a free, template-based video editor, not a standalone AI video product. It's a strong fit for creators who already edit inside CapCut and want AI to handle captions, background removal, or a rough first draft from a script. It's a weaker fit for anyone who wants a finished video from a single instruction with no manual editing, full model choice, or a complete soundtrack rather than a bare voiceover. Read the June 2025 terms of service update carefully before syncing client or commercial footage to CapCut's cloud. For everyday editing inside the app, this CapCut tutorial is a good starting point; for a different category of tool built around finished output rather than manual assembly, pexo.ai takes the description-to-finished-video approach instead.