✳ Kyan Chat

ONE WORKSPACE. YOUR CHOICE OF AI.

A little clarity.
A lot of possibility.

Sign in with your Kyan Barth account.
The same account you use on our other sites.

NameEmailPassword

Your API keys are encrypted on this server.

✳ Kyan Chat

YOUR CONVERSATIONS

K

Kyan Chat◉ Personal workspace

✳

A SPACE TO THINK

What’s on your mind?

Start with a question. See where it takes you.

Connect a provider in settingsAI can make mistakes. Check important information.

YOUR WORKSPACE

Providers & models

Connections and selected models are saved to your account.

Add a connection

Connection nameChat endpointAPI formatAPI key

Import a ready-made cURL command

Paste a provider cURL command with your API key. A GET /models request is enough. Models are fetched automatically; any example model or messages in the command are ignored. Choose your model next to the chat input.

Connect your own Ollama

Install Ollama and download a model on this computer. Allow https://gpt.kyanbarth.com in OLLAMA_ORIGINS, then restart Ollama. Keep Ollama bound to localhost.

Setup commands

Stop an already-running Ollama instance before starting it with these settings.

Terminal (Linux / macOS)

OLLAMA_ORIGINS=https://gpt.kyanbarth.com ollama serve

Windows PowerShell

$env:OLLAMA_ORIGINS="https://gpt.kyanbarth.com"
ollama serve

For the desktop app or a system service, set the environment variable in its settings and restart it. Ollama setup guide

Allow local network access if your browser asks. Some browsers block local HTTP connections from HTTPS websites; use a browser that supports local network access. Inference runs on your computer; conversation history is saved to your Kyan Chat account.