Central
Server Manager
+ KAL-AI Instance
+ Custom Instance
🧠 + Ollama Server
Refresh
Region:
ap-south-1 (Mumbai)
us-east-1 (N. Virginia)
us-west-2 (Oregon)
eu-west-1 (Ireland)
ap-southeast-1 (Singapore)
KAL-AI Instances
0
Loading...
Registered Servers
0
Loading...
Other Central Instances
0
Loading...
🧠 Ollama LLM Servers
Private / No Data Leaves VPC
0
Loading...
Launch Instance
🌐 Domain Assignment (ruvvy.com)
Subdomain Name
*
.ruvvy.com
Lowercase letters, numbers and hyphens only. Must be unique.
Instance Name
Instance Type
AMI ID
▶
Advanced Settings
Key Pair Name
Security Group ID
Instance Profile (IAM Role)
— required for SSM
Subnet ID
(optional)
Save as defaults for future KAL-AI launches
🧠 Launch Ollama LLM Server
Runs on g4dn.xlarge (T4 GPU, 16 GB VRAM) · private VPC only · no data leaves AWS
Server Name
Model to Pull
Llama 3.1 8B — best balance (5 GB, ~50 tok/s)
Llama 3.2 3B — fastest (2 GB, ~90 tok/s)
Mistral 7B — good instruction following (4.5 GB)
Qwen 2.5 7B — strong multilingual (4.5 GB)
Llama 3.1 8B Q8 — higher quality (8.5 GB)
Gemma 2 9B — Google model (5.5 GB)
Model is pulled (~5 GB) on first boot — takes 5–10 min. Instance cost: ~$0.53/hr.
After launch:
1. Wait ~10 min for model to pull (check Logs tab)
2. Copy the Private IP shown in the table
3. Go to
golden.ruvvy.com → Config → Local Model
and paste it