Nebula brings outside AI agents into a shared team workspace

Why it matters

Nebula is selling a shared workspace around agents teams already use, while charging per seat and metering its own agents separately. The model makes existing subscriptions a route into the product, but the workspace still has to earn its place in teams' daily work.

Nebula, the AI workspace founded by Furqan Rydhan, used an October 2nd post to describe a product that puts teammates and AI agents in the same workspace, including agents users already run. The announcement on X says users can create agents or connect existing ones; Nebula's product site identifies Claude and Codex among the external agents it supports.

https://x.com/NebulaAI/status/2106102724798341286

Rydhan has built software companies across several waves of internet technology. His own biography lists him as AppLovin's founding CTO, a founder and CTO of Bebo, and a founder and CEO of thirdweb. He is also the founder of Founders, Inc., which lists Nebula in its portfolio and identifies Rydhan as the company's founder. Nebula applies a similar approach: teams direct software from a shared place, while agents take on pieces of the work.

Nebula's current materials describe channels where people can set a goal, discuss work and assign tasks to a roster of agents. The product documentation says Nebula can coordinate specialist agents for a channel, while users can add agents built by teammates or create one by describing its job. Claude Code, Codex and other coding agents can also be brought into the workspace. Nebula aims to coordinate agents from different builders inside a team's existing collaboration flow.

That separation reflects a common workflow: work often begins in a shared team conversation, while an AI task gets handed off into a separate chat, coding tool or automation service. Nebula is trying to keep the instructions, task handoffs and human oversight together. Its documentation describes controls for sharing agents across a workspace and for adding them to particular channels. It also says a user can ask Nebula to change a channel's agent team, with the system proposing the change and waiting for approval.

That coordination layer is also where the commercial test lies. Nebula's pricing page lists a $25 monthly charge per seat. Teams can connect external agents that continue to run on subscriptions they already pay for, while Nebula's own agents use credits. Nebula charges for the workspace and seats, then meters usage of its own agents separately. The model lets teams try the shared workspace with existing tools, while giving Nebula a route to earn more when customers use its agents.

For teams, this setup affects costs and control. Nebula's documentation says connected app accounts remain personal, and agents use the account of the person requesting work. Workspace members can also share API keys and integrations. Those rules shape whether a shared agent can act across a team's tools without turning access into an informal handoff of passwords or individual subscriptions.

Rydhan has described the underlying shift in direct terms. In an earlier post on X, he wrote that when he wants to build software, his mind turns to an agent doing work for him or his team, and that he struggles to see digital work remaining outside agents' reach. Nebula's shared-workspace design reflects that view: people give direction in a team setting, and agents operate within the same flow.

The October 2nd post establishes the feature pitch, not the productivity outcome. The product's success depends on whether teams find it easier to coordinate work in Nebula than to keep using their existing chat, coding and automation tools separately. Connecting outside agents could make adoption easier; it also means Nebula must prove that its workspace adds enough value to justify another per-person subscription.