A project is an open-ended AI chat assistant with a specific job. Unlike tasks, projects don’t end. You can have unlimited conversations, start new sessions whenever, and come back to old ones. The magic part: projects automatically pull in your task documents as context. So when you open a writing assistant, it already knows your brand voice, your offer, your audience. You don’t have to brief it every time.Documentation Index
Fetch the complete documentation index at: https://docs.worldbranding.com/llms.txt
Use this file to discover all available pages before exploring further.
Starting a chat
Click any project from the sidebar or Projects grid. You’ll land on the project’s overview page: a list of your past sessions plus a button to start a new one. Hit the button. The AI opens the conversation. Go.Choices
Some projects have choices, different “modes” you can pick before starting a session. For example, a content project might have choices for “Twitter post,” “LinkedIn post,” “Long-form article.” Each choice is its own version of the assistant with its own instructions and its own knowledge files. If a project has choices, you’ll see them as separate buttons on the project page. Pick the one that fits what you want to do. The conversation is scoped to that choice.Sessions
Every conversation is a session. Sessions are saved automatically and titled by the AI after the first reply. You can:- Open an old session and keep going where you left off
- Rename a session (right-click in the sidebar)
- Delete a session you don’t need
What the AI sees
Each project chat sends the AI:- The project’s instructions (set by the team)
- Knowledge files the team uploaded for that project
- Generated documents from any prerequisite tasks you’ve completed
- Documents you explicitly attached via the document picker
- The conversation history
Prerequisites
Just like tasks, projects can require certain tasks to be done first. If you see a lock screen when you open a project, that’s why. Finish the listed prereqs and the project unlocks. The reason: the project genuinely needs those documents to do good work. Skipping them means worse output.Voice + files in chat
Every chat has:- Voice input: hit the mic, talk, it transcribes via Whisper. Useful when you’d rather think out loud.
- File attach: drop in an image or document. The AI reads it as part of the current message.
- Document picker: paperclip icon. Attach docs from your library to the conversation as persistent context.