What is d5s?
Understand the d5s workspace and the four ways teams can turn context into governed work.
d5s is a workspace for doing real work with AI agents. It brings the conversation, the context the agent should remember, the tools it may use, and the record of what ran into one product.
An organization can govern several workspaces while each workspace remains a separate day-to-day and content boundary. Joining the organization does not automatically reveal every workspace: an organization Owner or Admin assigns access to a person or group.
The workspace switcher shows one organization's workspaces at a time. Settings uses the same hierarchy: organization administration comes first, followed by settings for the selected workspace.
The basic unit is not a disposable prompt. You can begin with a chat, keep useful context in a project, and then turn a stable pattern into an agent, automation, or dashboard without rebuilding the work from scratch.
Four ways to work
- Chats are interactive conversations. Use them to explore a problem, review intermediate work, or make decisions with the agent.
- Agents are persistent teammates. Each has a role, standing instructions, a home project, and one continuing conversation.
- Automations are repeatable agent jobs. They run manually, once at a chosen time, or on a recurring schedule, and each execution has its own run history.
- Dashboards are durable, agent-built outputs. The current result is easy to return to, while refresh runs and previous versions remain inspectable.
What stays consistent
Every chat, agent, automation, and dashboard belongs to a project under the hood. A project supplies reusable instructions, files, memory, skills, and connector activations. If you start a chat, automation, or dashboard without choosing a project, d5s creates an unlisted project for it automatically. A standalone chat can surface that context later with Save as project; automation and dashboard projects remain represented by those resources. An agent's home project is managed through the agent and intentionally hidden from project listings.
Work is executed in sessions and runs. The surrounding session keeps continuity; each run keeps its own progress, tool activity, usage, result, and failure state.
Start with the decision
Start a first chat if you want to learn by doing. If you already know the outcome you need, choose a work mode. To understand why the app is organized this way, read how d5s fits together.