Corevisor connects your browser to a language model running on your own hardware, such as a llama.cpp, Ollama, LM Studio or vLLM server at home, so you can chat with it from anywhere.

Can Corevisor read my chats?

No. Prompts and replies are encrypted end to end between your browser and your home server. Corevisor's servers handle sign-in and access control but never receive conversation content. Chat history is stored on your home server.

Do I need to open ports or set up a VPN?

No. The home server makes outbound connections only. Corevisor connects directly when your network allows it and falls back to a relay that only sees encrypted traffic.

Which models work with Corevisor?

Any model served through an OpenAI-compatible chat completions endpoint on your local network. You can add several models and choose between them in the chat.

Can I share my home LLM with other people?

Yes. Send a single-use invite link and choose which models each person can use. You can revoke access at any time.

Chat with your home LLM from anywhere

Corevisor connects your browser to the language model running on your own hardware. No port forwarding, no VPN, and your conversations are encrypted end to end, so they never pass through anyone else's servers in readable form.

Open Corevisor How it works

Your browserLaptop, phone, anywhere

End-to-end encrypted

Corevisor home serverRuns on your network

Local network

Your modelllama.cpp, Ollama, LM Studio, vLLM

How it works

Run the home server

Start the Corevisor home server on a machine that can reach your model. It only makes outbound connections, so there's nothing to open on your router.
Link it to your account

Approve the home server from your browser, then add your models by pointing Corevisor at their OpenAI-compatible endpoints.
Chat from any browser

Sign in on any device and start chatting. Your browser connects straight to your home server and your chat history stays there.

Built for people who run their own models

Private by design

Prompts and replies are encrypted between your browser and your home server. Corevisor handles sign-in and access control, never your conversations.

No networking setup

Corevisor connects directly when it can and falls back to a relay that only sees encrypted traffic. No port forwarding, dynamic DNS or VPN.

Works with your runtime

Use any model behind an OpenAI-compatible chat endpoint. Add several and switch between them from the chat.

Share with people you trust

Send an invite link to friends or family and choose which models each person can use. Revoke access whenever you like.

Images too

Attach a photo or screenshot when your model supports vision input.

Your hardware, your rules

Keep your model, your API keys and your chat history at home. Only your home server knows where your model lives.

Questions

What is Corevisor?: Corevisor connects your browser to a language model running on your own hardware, such as a llama.cpp, Ollama, LM Studio or vLLM server at home, so you can chat with it from anywhere.
Can Corevisor read my chats?: No. Prompts and replies are encrypted end to end between your browser and your home server. Corevisor's servers handle sign-in and access control but never receive conversation content. Chat history is stored on your home server.
Do I need to open ports or set up a VPN?: No. The home server makes outbound connections only. Corevisor connects directly when your network allows it and falls back to a relay that only sees encrypted traffic.
Which models work with Corevisor?: Any model served through an OpenAI-compatible chat completions endpoint on your local network. You can add several models and choose between them in the chat.
Can I share my home LLM with other people?: Yes. Send a single-use invite link and choose which models each person can use. You can revoke access at any time.

Your model is already running. Go use it.

Open Corevisor