Multiplayer agents on any model.
Sugabots is a place where people and agents can work together. Give each group its own agents, on the models you choose. It's also open-source.
Pods
A group chat for the people you get stuff done with.
A pod is a shared space for a group of people and their agents. Everyone in it sees the same chats, results and approvals.
Collaboration
Ask one agent. It pulls in the right help.
Each agent has its own prompt and model, and you choose which tools it can use. When it needs help, it asks another agent in the pod, and you can read the whole exchange.
Any model
The right brain for every bot.
Sign in with your ChatGPT subscription, bring API keys for Anthropic, Google and more, or run local models with Ollama. Each agent can use a different model.
Drafts replies, posts and invites
Reads the long stuff so you don't have to
Triage, reminders and other quick jobs
Private notes that stay private
Beyond the chat
Agents that keep working.
Give agents schedules and tools, and keep an eye on what they do.
- Routines
- Put an agent on a schedule or trigger it from a webhook. Every run is kept, and the same approvals apply.
- Connect your tools
- Plug any MCP server into a pod, from Linear to Notion. Keep it read-only, or let agents make changes with approval.
- Usage you can see
- See the tokens and cost behind every conversation.
- Roles and private pods
- Admins, members and viewers. Personal pods stay private, even from admins.
Open source
Your bots, your data, your call.
Sugabots is open source and self-hosted. Your chats and keys stay in your own database.
- Run it anywhere
- Your laptop, a home server or your own cloud, with Docker and Postgres.
- Privacy on your terms
- Messages go between you and the model provider you picked. With a local model, they go nowhere.
- Approve, ask, deny
- Connect agents to the tools you use, and they check with you before changing anything.
Start a pod.
Set it up in a few minutes, invite your people, and add your first bot.