Getting started

Providers

Servers you run, cloud providers with an API key, and plans you already pay for.

Draggy runs models on this computer by itself. Settings → Providers is where you add more: a server you already run, a cloud provider with an API key, or a ChatGPT, Claude or Google plan you already pay for. Each has its own group in that list.

On this computer

The Draggy engine comes first, with the models you downloaded in Settings → Models. Below it are the local servers Draggy knows: Ollama, LM Studio, llama.cpp server, vLLM, SGLang, Jan, LocalAI, KoboldCpp, text-generation-webui, GPT4All, llamafile, MLX-LM and Lemonade.

Start the server, then press Scan. Draggy tries each one's usual port on this computer and lists the models it finds. For a server on another port or another machine, set its Address under Advanced.

API keys

Press Add provider and pick one: OpenAI, Anthropic, Google Gemini, xAI, DeepSeek, Mistral, Groq, OpenRouter and a dozen more, or Custom for any OpenAI-compatible server. Get a key opens the provider's page for making one. Paste it into API key and press Test. When the key works, the provider's models are listed; tick the ones you want in the model menus.

The key is encrypted with your system's keychain before it is saved. On a system with no keychain, providers that need a key cannot be switched on.

Accounts

ChatGPT, Claude and Google take a sign-in instead of a key. Draggy reaches each plan through that company's own tool: Codex, Claude Code or the Gemini CLI, downloaded the first time you sign in. The sign-in page opens in your browser, and Draggy never sees your password.

Adding OpenAI, Anthropic or Google asks which way you want: Sign in with your plan or Use an API key.

Between releases

Draggy reads a list of new models and providers from draggy.org, so a provider can appear before the next version does. Adding one of those first shows the address your key and your conversations will go to.

What goes where

GroupWhere the model runsWhat it costs
On this computerThis computerNothing
API keysThe provider's serversEach request, billed by the provider
AccountsThe company's serversYour plan's limits

What each one sends, and to whom, is on the privacy page.