AI Assistant - Chat and Prompt Director

What is the AI Assistant

On the right side of every studio you'll find an AI Assistant — an artificial intelligence you can ask questions to receive ready-made prompts, content advice, and creative ideas.

The assistant has two modes:

• Agent — intelligence that reasons, generates, and assembles content (Claude models, more powerful)
• Direct chat — simple conversation with the model (fast responses, text only)

How to use the chat

1. Click on the Assistant section (or scroll the panel on the right)
2. Type a question or idea in the "Ask AI" field

• Examples: "give me 5 ideas for a LinkedIn post about AI", "create a prompt for a minimalist logo"

1. Press Enter or click Send
2. Read the response — the assistant suggests ready-to-use prompts or creative ideas

Specialization buttons

Below the chat you'll find buttons to focus the assistant on specific tasks. Depending on which studio is active:

• Reel — generate ideas for viral short videos (9:16)
• Hook — create engaging opening lines to capture attention
• Image Prompt — specialized for detailed visual prompts
• Video Prompt — scripts and descriptions for video generation
• Music Prompt — genre, mood, instruments for audio generation
• Social — hashtags, copy, tone of voice for social posts

Click a button and type an idea: the assistant will respond with output specific to that format.

Reusable history

The chat remembers the conversation while you're in the active studio. You can:

• Go back to previous messages to copy a prompt
• Continue the conversation — each new message sees the previous context
• Use the prompt in the studio — click "Use this prompt" on a response and the text goes directly to the studio's prompt field

Chat clears per section

When you switch from one studio to another, chat history stays isolated:

• Image Chat ≠ Video Chat ≠ Audio Chat
• Each studio has its own independent assistant
• Context doesn't mix between sections

This helps you keep conversations focused on what you're creating.

Choose your AI model

At the top right of the Assistant, you can select which AI to use:

• Claude Opus 4.8 — most powerful, reasons deeply (ideal for complex strategies)
• Claude Sonnet 4.6 — good balance between speed and quality
• Claude Haiku 4.5 — very fast, for simple responses
• Free (auto) — free cloud AI, picks the best available
• Local — AI on the Omnia Studio server
• Gemini — Google model

Some models may only be available to paid subscribers. If you see a lock 🔒, they're reserved for paid plans.

Tips

Be conversational. You don't need formal instructions. You can ask like you'd ask a colleague: "I need an idea for a post about AI that doesn't sound too technical."

Iterate with the AI. If the first response isn't perfect, respond with feedback: "Less generic, more specific to automotive" — the assistant will adapt the next prompt.

Copy and modify. The assistant isn't law: take the prompt, tweak it, add personal details. The best results come from collaboration.

Use specialization buttons. If you click "Reel" instead of typing "create a viral video", you'll get more precise output.

Common issues

Chat doesn't respond

• The chosen model might require login (Claude, Gemini). Connect your account in settings.
• If it's "Free (auto)", the server might be overloaded. Try again in a few seconds.

Error "model not available"

• Your administrator has disabled that model for your account.

Response isn't in English

• Add to the chat: "Respond in English". The assistant is multilingual and adapts.

"Use this prompt" button doesn't work

• Make sure you're in the right studio. The prompt is sent to the currently open studio.