Step 1 of 3
What's your name?
Vega will use this to personalise your experience.
Step 2 of 3
When's your birthday?
Helps Vega wish you well and tailor suggestions to your stage of life.
Step 3 of 3
What should Vega know about you?
Share anything useful — your job, hobbies, how you like responses, languages you speak. Vega will remember this.
Sign in to Vega
Get 50 prompts/day and access to all models
⏳
Daily limit reached
Your tokens reset at midnight.
Sign in for a larger daily allowance, or come back tomorrow.
⚠️
Delete your account?
This permanently deletes your account, all chats, and your profile. This cannot be undone.

Type delete to confirm.
Listening…
Speak now
Saved to memory
Notes mode — speak freely
Planning mode — Vega will autonomously search and build a plan for you
Incognito mode — this conversation won't be saved
BuzzCloud Vega  · AI
Code ✦
What's on your mind? I'm here to help.

Ask me anything — from life advice and meal ideas to quick questions and creative projects.

I can't stop overthinking
What should I make for dinner?
Help me reach out to an old friend
Home workout with no equipment
How do I sleep better?
Recommend me something to watch
↵ send ⇧↵ newline
Vega can make mistakes — always verify important information.
Modes
Response style
Voice
🧠 Saved Memory
0 items
Nothing saved yet. Vega will remember things automatically during your conversations.
Notes
0 words
Start speaking and your transcript will flow in here
Paste any text — meeting notes, article, transcript — and Vega will turn it into beautiful structured notes.
What's new ✦
Here's what we added
A few things you might have missed since your last visit.
Quiz Mode
Paste any text or pick a topic — Vega builds an interactive multiple-choice quiz with scoring, explanations, and difficulty settings.
Regenerate Response
Hover any AI message and hit ↺ to get a fresh answer. The old response is swapped out instantly.
Notes Mode
Speak freely — Vega transcribes in real time and compiles your rambling into beautiful structured notes with headers and bullet points.
Talk Mode
Have a full voice conversation with Vega. Improved voice quality — warmer, less robotic. Try it on the go.
Keyboard Shortcuts
Press ⌘/ to see all shortcuts. ⌘K focuses input, ⌘⇧N new chat, Esc closes everything.
Token-based Limits
Usage now scales dynamically by model. Bigger models cost more tokens — smaller ones go further on your daily budget.
Model Guide
All models run locally on BuzzCloud infrastructure.
✦ Lyra Series — Vision + Language
Lyra 1.1
👁 Vision⚡ Fast
4 billion parameters
Based on Qwen 3.5 4B by Alibaba Cloud
Lightweight multimodal model. Understands and reasons about images alongside text. Great for quick questions, image analysis, and low-latency responses.
Token cost: 3.5 chars = 1 token
Lyra 1.9
👁 Vision⚖ Balanced
9 billion parameters
Based on Qwen 3.5 9B by Alibaba Cloud
Mid-size vision model with significantly stronger reasoning. Handles complex image descriptions, charts, screenshots, and detailed visual analysis.
Token cost: 2 chars = 1 token
✦ Sulafat Series — Efficient Text
Sulafat 1.0
⚡ Fastest
3 billion parameters
Based on Qwen 2.5 3B by Alibaba Cloud
Smallest and fastest text model. Best for simple Q&A, quick lookups, and tasks where speed matters more than depth. Very light on resources.
Token cost: 4 chars = 1 token
Sulafat 1.7
⚖ Balanced
7 billion parameters
Based on Qwen 2.5 7B by Alibaba Cloud
A well-rounded everyday model. Good at writing, summarisation, coding help, and general reasoning. Solid choice for most tasks.
Token cost: 3 chars = 1 token
✦ Sheliak Series — Advanced Reasoning
Sheliak 2.0
🧠 Smart
8 billion parameters
Based on Qwen 3 8B by Alibaba Cloud
Newer architecture with improved instruction-following. Stronger at multi-step reasoning, code generation, and nuanced tasks compared to Sulafat series.
Token cost: 2 chars = 1 token
Sheliak 2.4
🧠 Smarter
14 billion parameters
Based on Qwen 2.5 14B by Alibaba Cloud
The most capable mid-tier model. Excellent at complex writing, analysis, coding, and long-context tasks. Recommended for demanding use cases.
Token cost: 1.5 chars = 1 token
✦ Vega — Flagship Code Model
Vega 3.0 ★ Flagship
🔥 Powerful🐢 Slower
30 billion parameters
Based on Qwen 3 Coder 30B by Alibaba Cloud
The heaviest model in the fleet. Optimised for coding, system design, and complex multi-step problems. Uses significantly more tokens per response.
Token cost: 1 char = 1 token