> For the complete documentation index, see [llms.txt](https://docs.meganova.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.meganova.ai/quickstart/deploy-products.md).

# Deploy Products

## Summary

There are 3 main products at MegaNova AI:

* [**Inference Models**](/inference-models.md) — OpenAI-compatible serverless endpoints for text, vision, image, video, audio, and embeddings. No infrastructure to manage; call a model and pay per token or per output.
* [**Compute Services**](/compute.md) — dedicated GPU compute for workloads that need pinned capacity, including serverless model hosting, dedicated endpoints, and GPU rentals across tiers (4090, A6000, L40S, A100, H100, H200).
* [**Character Studio**](/character-studio.md) — purpose-built platform for creating, testing, and deploying AI characters and companion apps.

## Inference Models

Endpoints that host generative AI models immediately available to use. The same base URL (`https://api.meganova.ai/v1`) exposes every modality, and the OpenAI SDK works with `base_url` + your API key.

See the full catalog in the [Model List](/inference-models/model-list.md) and the [Core Services overview](/inference-models.md) for the quickstart.

## Compute Services

When serverless doesn't fit — you need pinned latency, custom models, or guaranteed throughput — use [Compute Services](/compute.md) to rent GPUs by the minute or run [dedicated endpoints](/compute/dedicated-endpoints.md).

## Character Studio

[Character Studio](/character-studio.md) lets you author AI characters with persona, memory, lorebooks, and quality benchmarks, then embed them in your own product or deploy them as standalone chat widgets.