Projects

Projects in Lisa are more than folders for issues. They're contextual workspaces — each one carries its own AI instructions, knowledge base, and scoped chat sessions. When you talk to Lisa inside a project, she's already tuned to that project's goals, constraints, and technical context.

Think of a project as the container where strategy meets execution. You define the "what" and "why" in the project's description and instructions. Lisa uses that context to generate issues, artifacts, and plans that actually match how your team thinks about the work.

Creating a project

Create projects through the UI, AI chat, or API. Every project has:

When you create a project through chat, Lisa can set all of these fields based on your conversation — including generating the description and instructions for you.

Custom AI instructions

This is where projects get powerful. Each project can have custom instructions that modify how Lisa thinks and responds when working within that project's context. For example:

Focus on performance optimization. All issues should include
benchmarking criteria. Prefer PostgreSQL-specific solutions
over generic SQL. Code examples should use TypeScript.

When Lisa operates within this project — whether you're chatting, creating issues, or generating a PRD — these instructions are automatically included in her system prompt. This means you define your team's standards once, and Lisa applies them every time.

Project status

Projects don't have a manual status toggle. Instead, Lisa computes the status dynamically based on the issues linked to the project:

• Empty — no issues yet.
• Planning — issues exist but none are in progress.
• Ready — issues are scoped and ready to start.
• Active — at least one issue is in progress.
• Completed — all issues are done.
• Cancelled — all issues are cancelled.

This means your project status always reflects reality — no one needs to remember to update it.

Board view

Switch to board view at any time to see your project's issues organized as a kanban board. Issues are grouped into five columns — Backlog, To Do, In Progress, Review, and Done — and you can drag and drop to update status instantly.

Each card shows the issue type, priority, title, and number. Click any card to jump to the full issue detail.

Knowledge base

Projects have an attached knowledge base made up of artifacts — PRDs, specs, architecture documents, design docs, and notes. These appear in the project sidebar under "Knowledge."

Lisa references these artifacts when working within the project. If you've written a PRD for user authentication and then ask Lisa to create issues for it, she already knows the requirements because the PRD is part of the project's knowledge base.

Project sessions

Every conversation within a project creates its own session with full history. This means you can run multiple planning threads in parallel — one session for scoping the auth feature, another for breaking down the notification system — without context bleeding between them.

Past sessions appear in the project sidebar. Click any session to resume where you left off. Lisa retains the full conversation history, so you don't need to re-explain context.

The project detail page

When you open a project, you'll see:

• Sidebar with two tabs:
  • Context — edit the description and AI instructions, browse linked artifacts (knowledge), and see past chat sessions.
  • Issues — a quick list of all issues tagged to this project, with status badges.
• Main area — an embedded chat interface, scoped to this project. Everything you say here is informed by the project's instructions, knowledge, and linked issues.

This layout keeps planning and context side by side. You're never more than a glance away from the project's goals while you're creating the work to achieve them.