Features
Chat with agents and people, build your own agents, bring your whole team into one workspace, and pick the model that fits the job.
Chat
One Place to Talk to Anything
Talk to AI agents and real people in the same conversation. Mention an agent inline, message your team, send a voice note, or go hands-free with Voice Mode.
Mention agents inline
Pull any agent into the conversation with an @mention and get its answer right in the thread.
Direct and group messages
Message teammates one on one or in groups, with emoji reactions and audio messages.
Voice Mode, dictation, read aloud
Speak in real time, dictate instead of typing, and have replies read back in the voice you choose.
Attachments and image generation
Send images and files, generate images in chat, and organize conversations into projects.
Agents
Build an Agent That Works for You
Compose a custom AI agent with its own personality, expertise, tools, knowledge, and guardrails. Keep it private or share it in the Marketplace.
Define who it is
Set the name, intro message, system prompt, avatar, and voice persona.
Connect real tools
Web search, email, calendar, messaging, file management, REST APIs, and integrations like QuickBooks and Microsoft 365.
Ground it in facts
Give it procedural, declarative, and contextual knowledge so answers come from your material.
Set the guardrails
Decide what the agent will and will not do before it ever talks to anyone.
Organizations
One Workspace for the Whole Team
Invite your team, share agents and knowledge, and keep business context in one place. One workspace, one brain, the whole team.
Invite teammates
Add your team with roles and permissions, and run more than one organization from a single account.
Shared agent library
Build an agent once and the whole organization uses it. Update it once and every teammate gets the change.
Organization Brain
Upload documents to a shared knowledge base so agents answer with your company's facts and cite the source.
Shared context and controls
Organization profile, house style, policy guardrails, shared API keys, and usage analytics in one place.
Multi-Model LLM Picker
Pick the Model That Fits the Job
CubDen is not locked to one AI provider. Choose which model answers from the chat composer, and switch whenever a task calls for something different.
Anthropic
OpenAI
xAI
QwenModels from multiple providers
Frontier and open models side by side in one picker, each marked with its provider.
Always current
The list loads live, so new models show up as soon as they go online. No app update needed.
Same experience on every model
Streaming replies, agents, and tools work the same way whichever model you pick.
Sensible default
A default model is preselected, so you can start chatting right away and change it only when you want to.