Skip to main content
Enclave
English
Start free
Local models

Use self-hosted Hugging Face TGI in Enclave

Enclave supports Hugging Face TGI: point the model layer at an inference endpoint you deployed with TGI (OpenAI-compatible API), and characters reply with your private HF-ecosystem models. Ideal for advanced users and teams with GPUs who want inference fully private.

Last reviewed:

The short answer

Point Enclave's model layer at your self-hosted HF TGI (OpenAI-compatible endpoint) to drive characters with private inference.

Key points at a glance

  • Private inference: models run on your own GPU/server.
  • OpenAI-compatible API — configured much like connecting a cloud API.
  • Use the many open-source models on Hugging Face, with the deployment under your control.

How to connect

First spin up an OpenAI-compatible inference endpoint with TGI and load your chosen HF model, then in Enclave's model settings choose an OpenAI-compatible provider, point the base URL at your TGI address, enter the key (if any), and set it as default or assign per character.

Who it's for

Teams and advanced users with GPU resources, deep in the Hugging Face ecosystem, or who need fully private inference. With Enclave self-hosted, both models and data stay on your own infrastructure.

Key facts

Integration
OpenAI-compatible endpoint (base URL)
Billing
Self-hosted, no third-party API fees
Best for
Private inference / HF-ecosystem models
Data flow
Stays within your own infrastructure

Ready to try it?

Open it in your browser — no credit card, no install.

Start freeLearn about self-hosting

Related

Connect other models

Use self-hosted Hugging Face TGI in Enclave · Enclave