Your First Chat
After reading this page you will be able to: send your first message, get an AI reply, and handle any permission request.
1. Send your first message
Section titled “1. Send your first message”- Open the main window and Select an assistant to start a task on the welcome page (when unsure, pick a general-purpose assistant)
- Click the model selector above the input box and make sure HuiAI is selected; after completing Required Setup it is the default
- (Optional) Under “Work in a project”, click Select project folder so AI works inside a specific folder; if the task involves no files, keep No Project
- Type your request in the input box and press
Enterto send
2. Two ready-made prompts
Section titled “2. Two ready-made prompts”Paste either of these into the input box (pick the project folder first for the second one):
Look through this folder read-only. List the files it contains and what each is for. Do not modify anything.Build me a beautiful personal homepage and open it for me when it's done.The first one tests “reading”, the second one tests “building”. If both work, your setup is fine.
3. How to handle permission requests
Section titled “3. How to handle permission requests”Before AI reads files or runs commands, the window shows a permission request describing what it wants to do:
| Button | When to choose it |
|---|---|
| Yes, allow once | Allow this one time only; ask again next time (recommended on first use) |
| Yes, allow always | Trust this kind of operation; stop asking |
| No (esc) | Refuse when unsure or unwilling |
When in doubt, read the paths and commands in the request before deciding. Full rules are covered in Permissions.
4. If it did not work
Section titled “4. If it did not work”| Symptom | Check first |
|---|---|
| Token check fails | Re-save and re-check the token following Required Setup |
| Model errors | Check the model configuration on the Models page |
| Agent not ready | Check the agent status on the Agents page |
What success looks like
Section titled “What success looks like”- After the message is sent, AI starts replying
- When a permission request appears, you can read it and make a choice
- You receive a complete reply with no errors
