# Hugging Face — Remote VAEs for decoding with Inference Endpoints 🤗

- Company: Hugging Face (huggingface.co)
- Announced: 2025-02-24T00:00:00+00:00
- Category: capability-change
- Subject: Platform
- Models affected: SD v1.5, Flux, HunyuanVideo
- Source: https://huggingface.co/blog/remote_vae
- Record: https://forck.live/items/1741-remote-vaes-for-decoding-with-inference-endpoints

Hugging Face introduces a remote VAE decoding feature for Inference Endpoints, allowing users to offload the VAE decoder to a separate endpoint to reduce memory usage and improve latency, with support for models like SD v1.5, Flux, and HunyuanVideo.

## Evidence

Verbatim from https://huggingface.co/blog/remote_vae:

> we want to pilot an idea with the community — delegating the decoding process to a remote endpoint.

---

Record: https://forck.live/items/1741-remote-vaes-for-decoding-with-inference-endpoints
Catalogue: https://forck.live/llms.txt
Feed: https://forck.live/feed.md