Review

Captions Review

Captions is a strong buy for creators and teams that want fast AI video editing, localization, and avatar workflows, but its pricing structure and privacy tradeoffs make the decision narrower than the marketing suggests.

Last updated April 2026 · Pricing and features verified against official documentation

Captions used to be easy to file away as a captioning app. That is no longer accurate. The product now sits somewhere between a consumer video editor, a generative video system, and a localization layer for short-form content, which is a more ambitious position and a more useful one for the right buyer.

That evolution matters because the company has been unusually public about it. TechCrunch reported the move to freemium in January 2025, then the rebrand to Mirage in September 2025, and then a fresh funding round in March 2026. This is not a mature utility resting on one trick. It is a product line that keeps widening its surface area.

The honest case for Captions is that it is one of the quickest ways to turn talking-head footage into something publishable without making you live inside a traditional editor. If your work is short-form social video, promotional clips, sales content, or multilingual distribution, Captions compresses a lot of tedious work into one app.

The honest case against it is that the product is opinionated in all the ways that matter. It is built around social video, not broad video production. It is also priced and packaged in a way that rewards people who understand the plan ladder, the platform split, and the company’s rebrand. That is too much friction for a simple app, and too much constraint for a universal one.

What the Product Actually Is Now

Captions is best understood as Mirage’s creator-facing video product, not as a standalone captioning utility. The public product now spans AI editing, prompt-to-video, AI actors, dubbing, caption styling, and a separate developer API, with web, iOS, and Android access. The current positioning is closer to “AI video production in a phone app” than “subtitles for clips.”

That shift also explains the awkwardness around the name. The company is Mirage, the consumer app is Captions, and the enterprise offer is now expressed through the same family. If you are evaluating it casually, the branding can look like a relic. If you are buying it deliberately, the branding is a clue that the company wants to sell video generation, not just editing.

Strengths

It gets you from raw footage to postable video fast. AI Edit is the core reason to care here: you feed in footage and get a cut with scene changes, B-roll, sounds, and pacing adjustments without manually assembling everything yourself. TechCrunch’s coverage of the feature makes the limitation plain too: it is aimed at a narrow class of content, but within that class it is genuinely efficient.

Localization is not bolted on after the fact. Captions can generate subtitles in 100+ languages and dub voiceovers in 30+ languages, which makes it far more useful for international distribution than a tool that only handles English first and asks you to solve the rest elsewhere. For teams posting the same content across multiple markets, that matters more than another layer of visual polish.

The avatar and prompt workflows are practical, not just flashy. AI Twin, AI Creator, and the chat-based editor let you produce talking videos from prompts, scripts, or selfies with less production overhead than a normal editing flow. That will never replace a real shoot for every use case, but it does remove enough friction to change how often a small team can publish.

The platform has a real escalation path. The public pricing now stretches from a free plan to Pro, Max, Scale, and Enterprise, and the enterprise package adds training data exclusion, custom seats, and account management. That gives the product a believable story for both individual creators and organizations that care about volume and control.

Weaknesses

It is still a social-video tool wearing platform clothes. Captions is strongest on short, talking-head, vertical content. That is exactly why the AI edit flow works, but it also means the product is not trying to be a general-purpose editor. If you need timeline control, multi-cam complexity, or longer-form production, the app’s strengths become a constraint.

The plan structure is easy to misread. The headline pricing page reflects iOS plans only, while Android has a separate Lite tier and the help center notes paid plans can be monthly or yearly. That is normal for an app-store business, but it is not especially clean for buyers trying to compare options quickly.

The privacy story is acceptable only if you read it carefully. Mirage’s policy says it collects user-generated content, audio/video/text data, and some biometric information for feature support. It also says certain AI features require sharing your information with third-party partners, and that the company may disclose information to third parties for commercial purposes, including training their algorithms. Enterprise training-data exclusion is a real improvement, but it is not the default consumer posture.

Pricing

Pro at $9.99 per month is the default choice for most individual users. It covers the basics that make Captions worthwhile in the first place: captions, fast fixes, styling, and no-watermark exports. If you mainly want better-looking social videos without a lot of generative experimentation, Pro is the plan that makes sense.

Max at $24.99 per month is the tier for people who actually want the generative video side of the product. The extra credits, AI actors, chat-based editing, and custom B-roll matter if Captions is becoming your main production system rather than a convenience app. If you only dabble in those features, Max is easy to overbuy.

Scale is for volume, not novelty. At $69.99 per month and above, it is essentially a throughput package with more credits and more usage headroom. Scale 2x and Scale 4x are not meaningfully different product experiences; they are just capacity multipliers for teams already pushing the system hard.

The hidden trap is that the public pricing is not one coherent universal table. The page explicitly says the listed features and prices reflect iOS plans only, and the Android Lite option is separate. That is a fine packaging decision for the company, but it means buyers need to read the fine print before assuming a tier works the same everywhere.

Privacy

Captions does not present itself as a no-data, no-sharing product. The privacy policy says Mirage collects user content, account details, usage data, and some biometric information where needed for features like voice-to-text or augmented reality. It also says the company may use third-party partners to enable AI-based features, and that users direct Mirage to share information with those partners when they turn those features on.

The most important detail for professional users is what the policy does not promise by default. It does not read like a blanket consumer guarantee that your content is never used in ways that improve the service or feed partners. Instead, the policy allows disclosures to third parties for their own commercial purposes, and the terms say usage data may be used for business purposes tied to operating and improving the service. If that is too loose for your organization, the enterprise plan’s training-data exclusion is the version to buy, not the consumer tiers.

Who It’s Best For

• Short-form creators who publish often. Captions is a good fit if you make talking-head clips, promos, or social posts and want the app to do the repetitive edit work for you.
• Marketing teams that need localization. If you need the same video turned into multiple languages fast, the captioning and dubbing stack is more useful than a conventional editor.
• Brands testing AI-native video production. The AI actor, AI Twin, and prompt-to-video workflows make sense when you care about speed and iteration more than full manual control.
• Teams that need a growth path to enterprise controls. If you think this will spread from individual creators to a managed rollout, the enterprise package gives you a plausible upgrade path.

Who Should Look Elsewhere

• If you want a more traditional editor and stronger documentary-style workflow, look at Descript.
• If your priority is high-end generative video rather than social-first editing, compare Runway.
• If you want polished avatar-led video at scale, Synthesia is the more direct enterprise comparison.

Bottom Line

Captions is worth taking seriously because it solves a real bottleneck: getting talking video from rough capture to something publishable without moving through five different tools. That is more valuable than a novelty app and more specific than a generic AI assistant. For creators and marketing teams that live on short-form video, it has enough speed, localization, and generation depth to justify the subscription.

The price of that usefulness is narrowness. Captions is not trying to be every kind of video editor, and Mirage’s privacy and platform packaging reward careful reading rather than casual sign-up behavior. If you want a fast AI video system with a believable enterprise path, it is one of the stronger options. If you want a neutral editor, it is the wrong category.

Pricing and features verified against official documentation, April 2026.