What is Sugabots?
Sugabots is a shared workspace where teams work alongside bots and keep their conversations in one history. Teams can give bots different models, then follow tool calls and exchanges when bots pass work to one another. It is aimed at teams that want to collaborate with AI in a self-hosted setup, using their own model access, or eventually through a hosted version that is currently taking sign-ups.
Sugabots Features:
Shared Team History: People and bots work in the same team space, with a shared history that lets colleagues see what was asked and add context.
Model Choice: Teams can select a model for each bot, choosing different models for different jobs within the same team.
Visible Tool Calls: The workspace shows bot tool calls and their results so people can follow the work behind an exchange.
Bot Handoffs: When a bot needs help, it can ask another bot on the team, with the request, reply, and next steps visible.
Model Access: Bots can use a ChatGPT or SuperGrok subscription, API keys, or local models run with Ollama.
Self-Hosted Setup: The open-source app can run on a laptop or a team's own cloud using Docker, with model keys supplied by the team.
Sugabots Benefits:
Easier Continuity: A shared record helps teammates resume work without relying on someone to repeat the conversation or its background.
Clearer Oversight: Seeing bot requests and results makes it easier to understand how work moved between bots and what happened next.
Job-Specific Models: Choosing models per bot lets a team assign different model options to different kinds of work.
Team Collaboration: Keeping people and bots in one workspace makes it easier for colleagues to add context and continue a task together.
Deployment Choice: Teams can start on a laptop or their own cloud rather than waiting for the hosted version to open.
Use Cases:
Customer Reply Drafting: A sales or account team can ask a bot to draft a response to a customer's pricing question, then review it in the shared thread.
Pricing and Margins: A team can ask a bot to check whether a proposed discount still meets a margin target before preparing a quote.
Review Before Sending: A colleague can leave a drafted customer response in the conversation while another teammate holds it until a meeting is complete.
Bot-to-Bot Research: A team can follow a bot's request for help from another bot and inspect the response before continuing the work.
Local Team Workspace: A team that wants to run its own setup can install Sugabots on a laptop or cloud server and connect its model access.