Skip to content
Pre-releaseGantry has not shipped yet. These pages describe the designed behaviour and will be revised with the first release.

Projects

A project is a home for related chats. It holds:

  • Folders the agent may work in. A chat in the project can read, search and (depending on the permission mode) change files in them.
  • Knowledge files: documents you want every chat in the project to know about, sent as context.
  • Instructions: a short text added to the system prompt of every chat in the project, for conventions, tone or rules.
  • Connectors that are available to every chat in the project without attaching them one by one.
  • Artifacts produced by its chats, listed in one place.

Choose New project in the sidebar, give it a name, and add at least one folder if you want the agent to code in it. Chats created inside the project inherit its settings; a chat can still override the model and the permission mode.

Gantry’s system prompt has a fixed core that is never replaced, plus additive layers: your global instructions, the project’s instructions, and the chat’s own. Each layer has a size limit so the prompt stays predictable. Pinned skills are added the same way.