Skip to main content

Projects

Projects help you organize related work and provide context to your AI agents. When a project is active, its background information and files become available to the AI, enabling more relevant and informed responses.

What Projects Provide​

A project bundles together everything related to a specific initiative, client, or workflow. This typically includes context information describing what the project is about, files containing relevant documents and reference materials, and visual branding like colors and icons that help you quickly identify projects in the interface.

When you start a conversation with an active project, the AI receives that project's context automatically—no need to re-explain background information every time.

Creating a Project​

Navigate to Projects in the sidebar and click Create Project. Fill in the details:

Name should be clear and recognizable. You'll see this in dropdowns and lists, so make it easy to distinguish from other projects.

Description provides a brief summary shown in project lists—helpful for quickly identifying the right project.

Color and Icon give visual identity. Colors appear as a bar next to project-related items, and icons display in dropdowns and lists. Choose combinations that are meaningful to you or your team.

Project Context​

Context is information the AI receives when this project is active. Think of it as background knowledge that shapes how the AI approaches conversations.

Short Context​

A condensed description used for quick reference. This appears in project lists and gives at a glance information about what the project involves.

Long Context​

The detailed background information provided to the AI. This is where you put everything the AI should know when working on this project:

  • Project goals and objectives
  • Key stakeholders or client information
  • Important constraints or requirements
  • Background history and decisions made
  • Terminology specific to this project

Good long context helps the AI give relevant answers without you having to provide the same background repeatedly.

Working with Projects​

Activating a Project​

Click a project in the sidebar or use the project dropdown in the navigation bar. The active project's color appears in the interface, and its context flows into new conversations.

You can switch projects anytime—the new project's context applies to subsequent conversations.

Context in Conversations​

When you start a chat with an active project, the AI receives the project's long context as background information. This means the AI knows about the project's goals, constraints, and specifics before you ask your first question.

If you mention needing to check project files, the AI can reference documents you've added to that project.

Pinning Projects​

Pin frequently-used projects by clicking the pin icon. Pinned projects appear in your shortcuts for quick access, saving time when you switch between the same few projects regularly.

Managing Project Files​

Projects can contain files that agents reference during conversations. Upload relevant documents—specifications, reference materials, previous deliverables—and they become part of the project's knowledge.

To manage files, open the project details and use the files section to upload, view, and remove documents. See Project Files for more details.

Project Organization​

Editing Projects​

Click Edit on any project to modify its name, description, context, branding, or files. Changes take effect immediately—active conversations with that project will have access to updated context.

Archiving​

When a project is complete or inactive, archive it rather than deleting. Archived projects move to a separate list, keeping your main view clean while preserving the history. You can unarchive projects if work resumes.

Example: Setting Up a Client Project​

Here's how you might configure a project for client work:

Name: Acme Corp Website Redesign

Short Context: Website redesign project for B2B software company

Long Context:

Acme Corp is a B2B software company serving the manufacturing industry. They're redesigning their website to improve lead generation, provide a better mobile experience, update their brand presentation, and integrate with their Salesforce CRM.

Key contacts: Sarah (Marketing Director) is the primary stakeholder; Tom (IT Manager) handles technical requirements.

Timeline: Design phase runs through Q1, development begins Q2. Soft launch planned for end of Q2.

Brand guidelines: Modern, professional, blue/gray color scheme. Avoid stock photography; prefer custom illustrations.

The project includes homepage redesign, product pages, a new blog section, and a resource library for gated content.

Files: Brand guidelines PDF, current site analytics report, competitor analysis, wireframes

With this context in place, any conversation about this project starts with the AI already understanding who Acme is, what they need, and what constraints apply.

Tips for Effective Projects​

Be specific in your context. Vague context produces vague results. Include concrete details about goals, constraints, and important background.

Update context as projects evolve. If important decisions are made or requirements change, update the long context so future conversations reflect current reality.

Use files for reference material. Complex specifications, brand guidelines, or technical documentation work better as uploaded files than as inline context—the AI can search and reference them as needed.

Choose distinctive colors. When you have many projects, visual differentiation helps. Assign colors intentionally—perhaps green for active client work, blue for internal projects, orange for urgent items.