Chat
✦
How can I help?
Ask anything, run code, research, vibe code
Vibe Code
Describe what you want to build — Urai writes, runs, and iterates.
Knowledge Wiki
3-tier memory: search, add, and link knowledge pages.
Named Agents
Persistent AI agents with their own memory, tools, and schedule.
API Keys
Generate keys for apps, scripts, and integrations. Master key required to manage keys.
Generate new key
Active keys
Code Interpreter
Run Python, JavaScript, or Bash code — Docker-sandboxed when available.
(output will appear here)
Goals Agent
Create long-horizon goals. Urai plans steps and executes them in the background.
Loading…
Settings
Connection
Server URL
Urai API endpoint
API Key
Your urai_… token
OpenAI-compatible API
Point any OpenAI client at Urai — no cloud required.
Base URL
API Key
your urai_… key (or master key)Model
urai or urai-fast · urai-ultra · urai-cyberCode examples
Models
Local GGUF models served via llama.cpp. API-only models use your cloud keys.
Loading…
Ollama / GGUF import
Import any Urai GGUF into Ollama (Android app or desktop) for fully offline use.
Loading…
Import a custom GGUF into Ollama
# 1. Download any GGUF from the models list above # 2. Create a Modelfile: echo 'FROM ./your-model.gguf' > Modelfile echo 'PARAMETER num_ctx 4096' >> Modelfile # 3. Import: ollama create my-model -f Modelfile # 4. Run: ollama run my-model
On Android: use the Ollama Android app → import GGUF from storage → same Modelfile approach via termux or adb.
Generation
Max tokens
Default 512 · No-limit 8192
Integrations
Connect Google Calendar/Gmail or Microsoft Outlook. OAuth tokens stored locally.
Google Calendar & Gmail
○ Not connected
Microsoft Outlook
○ Not connected
Rate limit status
Loading…