HeyGen vs Captions (Mirage): Enterprise Avatars vs Short-Form Native
Both sell AI presenters and both do translation, but they were built for different rooms. Plan prices, the per-minute API rates, and which one fits a creator versus a company.
HeyGen and Captions (the app; the company renamed to Mirage in September 2025) overlap on avatars and on translation, and diverge on almost everything about who they are for.
Plans
| HeyGen | Captions / Mirage | |
|---|---|---|
| Free | yes | yes, since January 2025 |
| Entry | $29/month (Creator), ~600 credits | from ~$9.99, ~1,400 credits |
| Mid | $49/month (Pro), ~1,000 credits, 4K | Scale 2× (~2,800 credits) |
| Upper | $149/month (Business), ~1,500 credits | Scale 4× (~5,600 credits) |
| Enterprise | sales-led | direct quote |
Captions' reported figures vary more across sources than HeyGen's; check the current page.
The API rates are the clearest comparison
HeyGen translation on fal: fal-ai/heygen/v2/translate/precision at $0.05 per second — $3.00 per minute, read live on 2 August 2026. Up to 8 minutes, 175+ languages, one call covering translation, timbre preservation and lip-sync. precision and speed cost the same, so always use precision.
Mirage API bills by input duration rounded up to the minute, at a reported $0.45 per minute — a 2.5-minute video bills as 3.
Those numbers are not comparable head to head because they buy different things: HeyGen's is full translation with lip-sync, Mirage's is the captioning and editing pipeline. But both are unusually transparent for this category, and both are the cheaper path if your volume is modest.
What HeyGen is better at
Translation, decisively. 175+ languages in a single call with voice-timbre preservation and matched lip movement is the product other tools are compared against. On consumer plans it consumes about 5 credits per minute with lip-sync and about 2 for audio-only dubbing.
Enterprise fit. 4K export, collaboration, security controls, sales-led enterprise. If the buyer is a company with a compliance review, HeyGen is built for that conversation.
What Captions is better at
Short-form as a native format. Caption timing and styling, vertical-first workflow, the things the original app was built around and the new foundation models are being trained for. General-purpose tools approximate this; Captions does it properly.
AI Twin. A personalised avatar of you, rather than a stock actor from a library. Different offer from most of the category, and the right one if your face is the brand.
Price at entry. Roughly $10 against $29.
The reputation gap
Worth naming because it is asymmetric. HeyGen's product quality and its customer experience are rated very differently — its Trustpilot score sits around 2.4, with recurring complaints about billing and about plan terms being revised, including an "unlimited" tier being quietly capped. That is not a statement about the translation, which is excellent. It is a reason to prefer per-minute API access over a long commitment.
Mirage raised $75M in March 2026 to build its own models, which is a signal about direction rather than about current quality.
Which
HeyGen if translation is the job, or if you are buying for a company. Nothing else does multilingual dubbing this well in one call.
Captions if you are a creator publishing short-form daily and want captions plus your own avatar in one place at a lower entry price.
Neither, if what you want is one specific clip with your face in a scene you did not film. That is generation rather than presentation, priced per clip — ours starts at $0.99 — and both of these are better than us at everything downstream of already having footage.