Sumus AI has opened its collaborative workspace to the public, allowing people and AI teammates to work together in shared conversations. The platform, now available at Sumus AI, enables users to choose from multiple AI models for cloud agents, connect local agents, and invite colleagues, clients, friends, or family into the same dialogue. This approach matters because it moves beyond single-user AI interactions to a model where the person with the brief, the agent doing research, and the colleague reviewing results all participate in one space.
The core idea is to put the models people prefer and the people they work with on the same team. Users can assign supported models such as Grok, Claude, GPT, Gemini, and DeepSeek to different cloud agents. For example, someone might use Claude for drafting, a GPT agent to review arguments, and a Grok agent for another part of a brief. Roles are fully customizable, so a team could include a researcher, a writer, and a reviewer, each with distinct instructions, all working in a shared conversation where work is assigned and results discussed. Available models and usage depend on the chosen plan.
Collaboration extends beyond mere sharing. In Sumus AI, the person who configured an agent can invite others into a space and add them to the conversation, allowing teammates to ask follow-up questions, supply missing details, or request revisions without rebuilding the agent's setup. This differs from tools like Grok Bot, which offers shareable templates that give recipients a copy on their own account without the original conversation history. Sumus AI keeps everyone working with the same agent in the same conversation, so a colleague can query the assistant directly while the creator contributes expertise without becoming a go-between for every request.
The platform also connects supported local agents, including Codex, Claude Code, and Grok Build, with cloud agents and people. A user who already runs Codex or Claude Code on their computer can connect that signed-in agent, letting cloud agents handle research and planning while local agents manage implementation. Other participants can discuss the brief and review results in the conversation. The connected computer must remain online for local agents to work, though cloud agents continue independently.
Getting started involves creating a space, adding an agent, and inviting someone who will use or review the result. Participants can mention an agent to assign work, supply files, and respond to outputs, with the conversation evolving as the task develops. The Free plan includes 500 credits per month, up to three members, and up to ten AI teammates, with no credit card required. For larger teams, Plus offers 3,600 monthly credits, up to 25 members, and up to 300 AI teammates for $24 per month or $240 per year. The Sumus AI pricing page lists plan details and optional monthly credit add-ons; monthly credits reset and do not roll over.
By combining model choice, shared conversations, and local agent integration, Sumus AI addresses a growing need for AI tools that support group work rather than isolated prompts. Its public launch could encourage more teams to treat AI as a collaborative teammate, potentially changing how projects are staffed and reviewed across industries.
