Model Catalog

Browse and search the full catalog of Skytells inference models — filter by type, provider, and capability. Open any model to view pricing, SDK snippets, and input schema.

The Model Catalog is your browsable inventory of every inference model available on Skytells — covering Skytells' own families (TrueFusion, BeatFusion, Mera) as well as third-party models from Black Forest Labs, OpenAI, Google, Nvidia, and others.

In the Skytells Console, navigate to AI → Models in the left sidebar.

For the complete model list with namespaces, schemas, and pricing, see the Model Catalog API reference.

Catalog layout

The catalog is a filterable, scrollable table with one row per model. The top of the page has a search bar and filter tabs.

Search

The search bar filters models by name, namespace, or provider in real time. Searching flux returns all Flux variants; searching truefusion returns the full TrueFusion family.

Filter tabs

Above the table, a set of filter tabs narrows the list by model type:

Tab	Models included
All	Every available model, no filter applied.
Text	Language models — chat, reasoning, code, and instruction-following.
Code	Models optimized for code generation, completion, and analysis.
Image	Models that generate or edit images from text or reference images.
Video	Models that generate video from text or image inputs.
Audio	Music generation, speech synthesis, and audio editing models.
Music	Models dedicated to full music track generation.
Multimodal	Models that accept or produce more than one media type.
Embeddings	Models that produce vector embeddings for similarity and search.
All providers	Dropdown to filter by a specific provider (Skytells, OpenAI, Google, Nvidia, etc.).

Column picker

The Columns dropdown on the far right lets you show or hide any of the 11 available columns. Useful for reducing visual clutter when you only care about pricing or capabilities.

Table columns

Column	What it shows
Model	Model name and provider, with a provider logo. e.g., `Skytells / TrueFusion`, `Black Forest Labs / Flux 2 Flex`.
Context	Maximum context window (for text/LLMs). `—` for non-text models.
Latency	Typical latency class for this model.
Throughput	Relative throughput rating.
Input	Pricing per input unit (token, image, second, GPU-hour, or prediction).
Output	Pricing per output unit. For image models, input and output pricing are often identical.
Capabilities	Visual badges showing what the model can do. Hover to expand.
Providers	Which provider serves this model — shown as a provider logo badge.
Zero Data Retention	Whether the provider guarantees zero data retention on inference requests.
No Prompt Training	Whether the provider guarantees that your prompts are not used to train or fine-tune models.

Reading pricing

USD 0.03/image — $0.03 per generated image.
USD 0.50 / 1M input tokens — $0.50 per one million input tokens.
USD 0.196/second — $0.196 per second of generated video.
USD 1.12/prediction — flat rate per prediction, regardless of output size.

For the full model pricing reference, see the Model Catalog API docs.

For the complete model list with namespaces, pricing, and schemas — including the full TrueFusion family, BeatFusion, Mera, and all third-party models — see List Models and the Model Catalog API.

From catalog to integration

Test first — Click any model row to open its Model Card, then select Try in Playground. The Inference Playground opens pre-loaded with that model.

Integrate directly — Copy the model namespace shown on the card (e.g., skytells/truefusion-pro) and pass it to the Predictions API or the TypeScript SDK.

Model Card — single-model detail view, opened by clicking any catalog row.
Inference Playground — test a model interactively before integrating.
Predictions — review the history of all inference requests.
API: List Models — fetch the catalog programmatically.
API: Model Catalog — full model reference with schemas and pricing.

How is this guide?

On this page