Projects: Organized Workspaces with Persistent Context
Each project holds conversations, files, custom instructions, and memory. Start a new conversation in a project and every AI already knows your context. No more re-explaining.
One initiative, one workspace. Your marketing strategy project doesn’t mix with your product roadmap project. Focus stays focused.
Starting every conversation from zero
You’ve had 20 conversations about your product launch. You start conversation #21 and have to explain the background again. “We’re a B2B SaaS company targeting mid-market, our main competitor is X, we’re launching in Q2…”
Context gets lost between conversations. Files you uploaded yesterday aren’t available today. Decisions from last week’s session are forgotten. You spend more time setting up context than getting value.
Projects solve this. Create a project, describe it once, and every conversation in that project starts with full context. The AIs remember your files, your constraints, your decisions.
Everything connected. Nothing lost.
Each project is a complete workspace for one initiative.
Conversations
All your chats within this project. Searchable, organized, and contextually connected. Each conversation benefits from the project’s shared knowledge.
Custom Instructions
Persistent rules all AIs follow. Define your context, constraints, audience, and preferences once. They apply to every conversation automatically.
Files
Upload documents for AI reference. Every conversation in the project can access them. No re-uploading, no lost context.
Memory
What the AIs remember from past conversations. Key decisions, important insights, and context that persists across sessions.
Knowledge Graph
Entities and relationships extracted from your work. The AIs build a structured understanding of your domain over time.
Isolation
Projects don’t leak into each other. Your marketing research stays separate from your product roadmap. Focus remains sharp.
Tell the AIs who you are – once
Custom instructions are persistent rules that every AI reads before responding. Write them once, benefit in every conversation.
Example: Product Development Project
We’re building a mobile fitness app for busy professionals (25-45).
Tech stack: React Native, Node.js, PostgreSQL, AWS.
Current stage: MVP with 500 beta users.
Competitor set: Peloton, Nike Training Club, Freeletics.
Key constraint: 2-person dev team, 6-month runway.
Example: Content Marketing Project
Brand voice: Professional but approachable. Never corporate-speak.
Target audience: Technical decision-makers (CTOs, VPs of Engineering).
Content goal: Thought leadership that drives inbound demo requests.
Topics we own: Developer productivity, AI-assisted workflows, team scaling.
Avoid: Generic advice, content that sounds like everyone else’s blog.
Every conversation in these projects starts with this context. The AIs never forget who you are or what you’re working on.
Master Projects: Cross-Project Intelligence
Regular projects are isolated – their knowledge stays within. A Master Project breaks that boundary. It can draw on knowledge from all your other projects.
Use a Master Project when you need to ask questions that span your entire body of work. Strategic planning that considers Product, Marketing, Sales, and Engineering perspectives. Quarterly reviews that synthesize progress across all initiatives. Pattern recognition across multiple projects.
Example: “Based on what we’ve discussed across all my projects, what are the three biggest risks to our company right now?” The AIs pull from Product (technical debt), Marketing (competitive pressure), and Sales (pipeline concerns) – synthesizing a cross-project view.
Upload once. Reference everywhere.
Add relevant documents to your project. Every AI can access them in every conversation.
Research Documents
Market research, competitive analysis, industry reports
Specifications
PRDs, technical specs, requirements documents
Strategy Docs
Business plans, pitch decks, strategic frameworks
Reference Material
Style guides, brand guidelines, process documentation
File limits by plan: 5 (Spark), 25 (Pro), 100 (Frontier), Unlimited (Enterprise)
Getting the most from Projects
One initiative per project
Don’t mix unrelated work. “Q1 Marketing Strategy” is good. “Everything about my company” is too broad. The tighter the focus, the better the AI responses.
Spend 60 seconds on the description
The project description becomes context for every AI in every conversation. A good description pays dividends across dozens of sessions.
Use clear naming conventions
“Q1 2026 Marketing Strategy” beats “Marketing Stuff”. Future you will thank you when you have 20 projects in your sidebar.
Start a new project when the topic changes
If you’re working on a fundamentally different initiative, create a new project. This keeps AI responses focused and prevents context pollution.
Context that persists. Focus that stays sharp.
Stop re-explaining your background in every conversation. Start a project and let the AIs remember.