Modellix review 2026: honest scoring on Aurora Mobile's unified AI media API, PixVerse V6 video, pricing, and what it means for creators versus developers.
Judged as what it is — a unified API to hundreds of AI media models — Modellix is a strong, well-built infrastructure product: one request format, transparent per-second pricing, a CLI, MCP, and a Playground, and a fresh, genuinely capable video model in PixVerse V6. The catch is scope. It returns a raw file and stops there — no captions, no formats beyond the model's output, no brand layer, and no publishing. Excellent developer plumbing; not a creator's content tool.
Aurora Mobile (NASDAQ: JG) added PixVerse V6 to Modellix on August 12, 2026, and that is the news that puts the platform on most creators' radar for the first time. So this review does two things: it scores Modellix as a product, and it is honest about who that product is actually for. Modellix launched April 9, 2026 as a unified gateway to third-party AI media models — an infrastructure play, not a consumer studio.
I build a content engine for a living, so I judge these tools with a working operator's eye. The short version: Modellix is very good at the specific thing it is built for. Aggregating more than 200 image, video, and audio models behind one standardized API, with public per-second pricing and unified logs, is a real and useful piece of plumbing — and PixVerse V6 through it is a legitimately strong video model, up to 15 seconds of 1080p with synchronized native audio in a single request.
The honest catch is the same one every model-aggregation product runs into, only more so here because Modellix has no consumer layer at all: it makes the asset and stops. There is no editor, no caption tool, no per-platform reframing, no persona or brand-voice governance, and no scheduler or publishing to any social network. That is not a bug — it is a scope decision — but it is the single most important thing to understand before you decide Modellix is your content tool rather than a component in someone else's build.
Everything below reflects Modellix's described state as of 2026-08-12. Aurora Mobile rotates the model roster and per-second prices, so reconcile specifics against modellix.ai before relying on them.
Modellix is Aurora Mobile's unified AI media generation platform. Rather than train a model, it aggregates more than 200 third-party image, video, and audio engines — Seedance, Kling, Veo, Hailuo, Imagen, Nano Banana, and now PixVerse V6 among them — behind one standardized API that normalizes request structures and simplifies nested model parameters. You reach it through the API, the modellix-cli command-line tool, an MCP connection, or a browser-based Playground, and you pay public per-second rates with unified call logs and audit trails. PixVerse V6, released March 30, 2026 and added to Modellix on August 12, 2026, is the flagship reason to look now: up to a 15-second, 1080p clip in a single request with synchronized native audio, precise camera control across shot types, multi-shot composition, and five workflows (text-to-video, image-to-video, start-and-end-frame transitions, video extension, and reference-to-video with up to seven tagged references). Modellix is developer infrastructure — its output is a rendered file, and everything after that is up to whoever calls it.
Modellix fits developers and enterprises embedding AI media generation into their own products: teams who want one integration and one billing relationship across many models, who value per-second cost transparency, and who will handle captioning, formatting, and distribution themselves. The Playground makes it a low-friction way to test a model like PixVerse V6 before wiring it into production. It is the wrong primary tool for a creator, marketer, or agency whose deliverable is finished, published posts — because Modellix returns a raw clip and leaves the finishing and publishing to you, which for a non-engineer means it does most of nothing on its own.
| Dimension | Score | Why |
|---|---|---|
| Model breadth & choice | 4.7 / 5 | One standardized API to 200+ image, video, and audio models, swappable behind the same request — best-in-class for aggregation. |
| PixVerse V6 video quality | 4.3 / 5 | Up to 15s of 1080p with synchronized native audio, camera control, and multi-shot in one request — a genuinely strong current-gen video model. |
| Developer experience (API/CLI/MCP) | 4.3 / 5 | Normalized requests, a CLI, and MCP support make it clean to integrate; it assumes an engineer, which is the point. |
| Pricing transparency | 4.4 / 5 | Public per-second pricing per model, with unified call logs and audit trails for cost control and debugging. |
| Playground / ease of testing | 3.8 / 5 | A browser Playground lets you try a model before coding, though it is a test surface, not a production workflow. |
| Reliability & infrastructure | 4.0 / 5 | Backed by a public company building it as a deliberate AI-infra product, with logging and audit trails; roster and prices do rotate. |
| Output finishing (captions, reframing) | 1.0 / 5 | None. Modellix returns a bare file — no captions, no per-platform aspect ratios, no hook overlay. Out of scope by design. |
| Non-video/image content formats | 1.0 / 5 | It calls media models; it does not produce carousels, quote cards, blogs, or newsletters, so one idea cannot become a content week. |
| Multi-platform publishing | 1.0 / 5 | No scheduler and no posting to any platform. Distribution is entirely on the caller. Out of scope by design. |
| Brand-voice governance | 1.5 / 5 | No Persona Brief, banned-word filter, or face-locked persona — nothing keeps output consistent across a body of work. |
Modellix's pricing is one of its stronger points, precisely because it is honest and legible. It bills usage per second per model at publicly listed rates, with unified call logs and audit trails so you can see exactly what each render cost. For a developer running AI media at volume, that transparency is genuinely valuable — it makes cost predictable and debuggable in a way that juggling multiple vendor SDKs and price sheets is not. The free Playground plus per-second billing also let you validate a model like PixVerse V6 cheaply before committing, and Aurora Mobile ran a 20% PixVerse V6 discount from August 12–25, 2026 to encourage exactly that.
The important caveat is that Modellix prices the render and only the render. That is the right model for infrastructure, but it means the total cost of turning a Modellix clip into a published post includes everything the price sheet does not cover: the engineering to caption and reframe it, the tooling to schedule and publish it, and the ongoing work of keeping it on brand. For a developer embedding generation into a product, per-second is fair and clear. For a creator comparing it to an all-in content tool, the sticker price understates the real cost of shipping finished content, because the finishing is a separate build.
Set against a content engine, the categories simply differ. Kompozy's credit-based plans (Starter at $99/mo, Pro at $299/mo, Enterprise custom) bundle generation across 18 formats plus multi-platform publishing into one price — you pay more per unit than a raw render, but the finishing and distribution are included rather than a project you staff.
| Use case | Fit | Why |
|---|---|---|
| Embedding AI video into your own app or product | Strong | This is exactly what Modellix is built for — one API to many models with per-second billing. |
| A/B testing many video and image models | Strong | Swapping PixVerse V6 for Veo, Kling, or Seedance behind one request format is a core competency. |
| Cost control on AI media at render volume | Strong | Public per-second pricing plus unified logs and audit trails make spend predictable and debuggable. |
| Quickly testing a model before committing | OK | The browser Playground is a low-friction test surface, though it is not a production workflow. |
| A creator who wants finished social posts | Weak | Modellix returns a raw file — no captions, no reframing, no publishing — so a non-engineer gets little usable on their own. |
| Running a multi-format content calendar | Weak | It generates single media assets, not carousels, blogs, newsletters, or a coordinated week of posts. |
| Keeping output on brand across a body of work | Weak | There is no persona or brand-voice governance layer of any kind. |
Modellix and Kompozy are not really competitors; they are neighbors on the same clip, and it is worth being precise about the seam. Modellix is where an AI asset is summoned — call PixVerse V6 (or 200 other models) through one API and get a rendered file back, cheaply and legibly. Kompozy is where that asset becomes a business: captions burned in, reframed per platform, hook overlays through HyperFrames, every output held to a Persona Brief, then scheduled and published across the eight social platforms plus blog and email with a per-post review gate and Autopilot.
Honestly, the two can chain — a developer could pull PixVerse V6 from Modellix and hand the file to Kompozy to finish and distribute. But for the creator, marketer, or agency who does not want to write code, Modellix is a layer too low: it is the plumbing, and Kompozy is the finished room. Judge Modellix as infrastructure and it earns its score; judge it as a content tool and you are asking it to do a job it was deliberately not built to do.
Modellix is Aurora Mobile's (NASDAQ: JG) unified AI media generation platform, launched April 9, 2026. It gives developers one standardized API — plus a CLI, MCP support, and a browser Playground — to more than 200 image, video, and audio models, including PixVerse V6, with public per-second pricing.
For a developer or enterprise embedding AI media generation into a product, yes — the model breadth, per-second transparency, and clean integration are strong. For a creator who wants finished, published posts, it is a poor fit, because Modellix returns a raw file with no captions, formats, brand layer, or publishing.
PixVerse's flagship video model, added to Modellix on August 12, 2026. It generates up to a 15-second, 1080p clip with synchronized native audio in a single request, with camera control and multi-shot composition, callable through the same standardized API as every other Modellix model.
No. Modellix returns a rendered file through its API, CLI, or Playground. It has no captions, no persona layer, and no scheduler or publishing. Finishing and distribution are a separate job for a content engine like Kompozy.
Effectively yes. Modellix is built around an API, a CLI (modellix-cli), and MCP, with a Playground for quick tests. Using it in a real workflow means writing against it, which is why creators who want a no-code path look to a content tool instead.
They are different categories. Modellix is a developer API for calling AI models; Kompozy is a no-code content engine that captions, brands, schedules, and publishes finished content across platforms. Use Modellix to summon an asset; use Kompozy to finish it and ship a week of posts from it.
Modellix bills usage per second per model at public rates, with a free Playground for testing. Aurora Mobile ran a 20% PixVerse V6 discount from August 12–25, 2026. Prices and the model roster rotate, so confirm current rates on modellix.ai.