Hexcoded is built for commercial video with a documented rights position: 30+ models on one credit balance, plus an actor library where human creators have passed a liveness check and earn on every render. It's the wrong choice if your work has to run offline, if you need an external post pipeline that touches performance, if you need Runway or PixVerse specifically, or if you need open weights to self-host.
We'd rather write this than have you find it out after signing up. Six situations where something else is the better answer.
What is Hexcoded actually for?
Commercial video where the rights position matters.
Two things distinguish it. A wide model roster on one credit balance, so choosing a model per shot doesn't mean maintaining four subscriptions. And an actor library with three sources — platform AI actors, your own, and human creators who submitted their likeness under a licence with a liveness check and face-match, and who earn every time a paying brand generates with them.
The second is the part that's hard to assemble elsewhere. Model access is a commodity; a documented consent trail isn't.
Which models does Hexcoded not carry?
Named plainly, because you'll find out anyway.
| Not carried | What's there instead | Does the gap cost you? |
|---|---|---|
| Runway | Nothing equivalent | Named in 2026 US perception rankings, so the name gap is real. A straight absence |
| PixVerse | Nothing equivalent | Also named in those rankings. A straight absence |
| Wan 3.0 | Wan 2.7 and Wan 2.6 | Alibaba's newer flagship is an application-gated beta with no independent measurement. Defensible, and we'd rather say it than let you notice |
| Open-weight models | Nothing — everything is hosted | Only matters if you need to self-host or fine-tune, which is out of scope for any hosted platform |
Source: Hexcoded roster as published on the homepage and pricing page, checked September 2026. Model availability changes — verify before relying on this.
Sora is deliberately absent and that one isn't a gap. The API shuts down on 24 September 2026 with no OpenAI replacement.
When should you use something else?
Six situations, and none of them is an edge case.
Your work has to run offline
Hexcoded's licence is digital only. Paid ads and organic posts on online platforms are covered, plus your own websites and apps. Television, print, outdoor and other offline use require a separate written agreement.
If a campaign is heading for broadcast or out-of-home from the start, that conversation has to happen before you generate rather than after. And if the answer is no, this isn't your platform for that job.
You need an external post pipeline
Anything containing an actor permits light edits only. Trimming, cropping, captions, music and end-cards are fine; re-generating, morphing or re-voicing a face outside Hexcoded is not — and it's listed as a prohibited use rather than merely a breach.
If your workflow depends on external performance work — a different read, a face adjustment in post, a recut delivery that alters the performance — that's a hard constraint rather than an inconvenience.
You need Runway or PixVerse specifically
If a client has standardised on one, or your existing workflow is built around a feature only one of them has, that's a real reason. No amount of roster breadth substitutes for the specific tool a pipeline already assumes.
You need open weights
If you need to self-host, run offline, or fine-tune on your own data, a hosted platform can't help. Worth being clear that this isn't a limitation anyone plans to fix — hosted access and self-hosting are different products.
You want one model and a simpler interface
Thirty-plus models is useful when you're choosing per shot and overhead when you're not. If your work is one format with one model, a single-model tool will be faster to learn and cheaper to run.
The roster earns its complexity on work with varied shot types. It doesn't earn it on a single repeated format.
You're at the lowest price point and need real people
The entry plan doesn't include real human actors or Talking Actors, and caps resolution at 480p with a core model set rather than the full roster. If your budget is at that level and your work needs a real face, another platform's entry tier may serve you better than upgrading here.
What's genuinely hard to get elsewhere?
One thing, and it isn't model access.
A documented consent trail. Human creators who passed a liveness check and a face-match at onboarding, whose likeness couldn't be published without clearing both, whose consent is a timestamped record kept permanently, and who earn on every render rather than once at the outset. When you cast, you're shown a consent-verified receipt.
That combination is a product decision rather than a technical capability, which is why roster breadth doesn't substitute for it and why competitors with better individual models don't necessarily have it.
Whether it matters depends entirely on your work. For a personal project it's overhead. For client delivery where someone has to approve the rights position, it's the whole thing.
Model access is a commodity. A consent trail isn't.
Where this falls short. A consent trail solves the likeness question and nothing else. It doesn't make a shot look better, it doesn't fix identity drift, and it doesn't help if the model you need isn't on the roster — or if your delivery is going to television.
How should you actually compare platforms?
Six questions that surface real differences rather than feature counts.
- Which models does it carry, and does that include the one your pipeline assumes?
- Do you own the output or hold a licence to it, and where is that output permitted to run?
- What can you do to the footage afterwards, especially where a real person appears?
- Does the platform train on your output by default, and where does the opt-out live?
- For any real face, what consent exists, what does it cover, and can you produce evidence later?
- Which law governs the terms, and where would a dispute be heard?
Notice that only the first is about models. The other five are about what happens after you've generated something, which is where most of the difference between platforms actually sits.
That last question is worth asking everywhere and rarely is. Hexcoded's terms and actor licence are governed by the laws of India, with jurisdiction in the courts at Gurugram, Haryana, and liability is capped at the amount paid in the month before a claim. None of that is unusual for a platform licence. It does mean the forum and the ceiling are defined in advance, and not where a US buyer would assume.
Model rosters, plan features and licence terms described here are current as of the publication date and change. Verify anything you're relying on for a purchase decision. Hexcoded's terms are governed by the laws of India, with jurisdiction in Gurugram, Haryana.
- Hexcoded fits commercial video where the rights position has to be documented. That's the case it's built for
- The licence is digital only. If your work has to reach television, print or out-of-home, that needs a separate agreement
- Light edits only where an actor appears. An external pipeline that touches performance is a hard constraint
- It doesn't carry Runway, PixVerse or Wan 3.0. If you need one of those specifically, look elsewhere
- If you need open weights or self-hosting, a hosted platform is the wrong category entirely
- One model and one repeated format is better served by a simpler single-model tool
- The entry plan has no real human actors and caps at 480p. Budget-level work needing a real face may fit elsewhere
- Compare on what happens after generation — ownership, scope, editing, training, consent evidence, governing law. Model lists are the easy part
It depends what you're missing. If you need a specific model it doesn't carry — Runway, PixVerse or Wan 3.0 — go to that vendor directly. If you need open weights to self-host, look at open-weight models rather than any hosted platform. If you want one model and a simpler interface, a single-model tool will be faster to learn.
No. Runway is named in 2026 US perception rankings, so the absence is worth knowing about. If your pipeline or your client assumes it, that's a real reason to look elsewhere.
Not under the standard licence. The scope is digital — paid ads and organic posts on online platforms, plus your own websites and apps. Television, print, outdoor and other offline use require a separate written agreement, so that's a conversation to have before you generate.
No. It's a hosted platform, and self-hosting is a different product rather than a missing feature. If you need to run offline or fine-tune on your own data, look at open-weight models.
A documented consent trail. Human creators who passed a liveness check and face-match, whose consent is a timestamped record kept permanently, and who earn on every render is a product decision rather than a technical capability — so a competitor with better individual models doesn't necessarily have it.
Ask six things: which models it carries, whether you own or licence the output and where it's permitted to run, what editing is allowed, whether it trains on your work by default, what consent exists for any real faces, and which law governs the terms. Only the first is about models.
See it before you decide
30+ models on one credit balance, over 1,000 human creators with consent verified, and a published licence you can read before signing up rather than after. Including the parts that don't flatter us.
Read the termsMore on tools, features and how the platform fits together in Product.