Your first chat
Send your first message, read the response metrics, and choose how 9xchat connects to models.
Once you have saved your UID, 9xchat opens on Chats and you can type straight away. New accounts start with $1 in welcome credits, which is enough to explore the workspace before deciding how you want to run it.
Send your first message
- Open Chats. A thread is already waiting, or choose New Chat.
- Type in the composer at the bottom, or pick one of the suggested prompts.
- Send it. The reply streams back with the model's cost, latency, and token counts shown alongside.

That per-message breakdown is worth noticing early: 9xchat shows what each exchange actually cost and how fast it was, rather than hiding usage behind a monthly bill.
Around the composer
- Model picker — change model per message, mid-conversation, including between a cloud model and a local one.
- Attachments — add a file to give the model something to work from.
- Voice — dictate instead of typing.
- Press / to jump straight into the composer.
Threads and tabs
Threads are listed on the left, grouped into folders, with ⌘F search and filters for all, starred, and archived. Search covers notes as well as chats.
A thread can hold more than one tab, so you can explore separate directions from the same conversation without losing either. Multiple tabs need Pro — see Credits, Pro, and usage.
Choose how you connect to models
9xchat gives you three ways to reach a model, and you can mix them freely.
9xchat credits
Pay as you go through 123sudo. Nothing external to set up — this is what your welcome credits use.
Your own API keys
Bring your own key for supported providers such as OpenAI, Anthropic, Gemini, or DeepSeek. You pay the provider directly.
Local models
Run a model on your own machine. Nothing leaves your device.
Set these up under Library → Models, where a provider stays marked Inactive until you add its key. See The Library for that screen, and Credits, Pro, and usage for who bills you in each case.
Make it yours
Two settings are worth visiting early, because they change every response you get afterwards:
- Your profile — tell 9xchat what to call you, what you do, and what traits you want the assistant to have. This context is applied globally instead of being re-explained in each chat.
- Default models — set under Settings, so new threads start on the model you actually want.