PromptBlocks Logo

PromptBlocks

Context That Builds Itself

You've set up prompts, global context, projects, and memories. Now let's see how they multiply each other's value. This is the full picture of PromptBlocks.

The Context Stack

When you send a message in a project chat, the AI receives:

Your Message
  + Your chosen prompt (if you used @mention)
  + Your communication style (if set)
  + Your assistant personality (if set)
  + Your Global Context
  + Project Description
  + Project File Names
  + Project Memories
  + Relevant file excerpts (if files were searched)
        ↓
    AI Response

Each layer adds value. Together, they create responses that are:

  • Personalized to you and how you work
  • Informed by your project context and files
  • Consistent with past decisions and preferences
  • Styled to match your communication needs

Switching Projects = Switching Context

Here's where it gets powerful: when you switch projects, your entire context shifts instantly.

Project A: Client Website Redesign

  • AI knows the brand guidelines
  • AI knows the target audience
  • AI remembers you chose the blue color scheme
  • AI speaks in professional-but-approachable tone

Project B: Personal Side Business

  • AI knows your business model
  • AI knows your constraints (limited time, bootstrap budget)
  • AI remembers you're targeting solopreneurs
  • Same AI, completely different context

No setup. No explaining. Just switch projects and keep working.

Full Scenarios by Role

Let's see the full context stack in action:

Small Business Owner

You: Run a landscaping company Global Context: Who you are, your communication preferences

Project: Commercial Expansion

  • Description: Goals, target market, constraints
  • Files: Competitor research, pricing analysis
  • Memories: Key decisions, stakeholder preferences

A typical conversation:

You: "Draft an email to pitch our services to the property management company we discussed."

The AI knows:

  • You're pursuing commercial contracts (project description)
  • The specific company you mentioned (memory from earlier chat)
  • Your professional-but-friendly tone (global context)
  • Competitive pricing info if relevant (from files)

You get a ready-to-send email without explaining any context.

Recruiter

You: Technical recruiter Global Context: Your specialization, timezone, communication style

Project: Series B Startup Search

  • Description: Company, role, requirements, compensation
  • Files: Job description, company culture doc
  • Memories: Candidate feedback, hiring manager preferences

A typical conversation:

You: "This candidate has 4 years of experience but strong distributed systems background. Write an outreach message."

The AI knows:

  • The role requires 5+ years (project description)
  • But the hiring manager values skills over tenure (memory)
  • The company's engineering culture and selling points (files)
  • How you like outreach messages structured (global context)

You get a personalized message that addresses the experience gap while highlighting fit.

Independent Consultant

You: Operations consultant Global Context: Your background, how you prefer structured thinking

Project: Manufacturing Client

  • Description: Engagement scope, deliverables, timeline
  • Files: Current process documentation, interview notes
  • Memories: Stakeholder dynamics, past recommendations, political context

A typical conversation:

You: "Help me structure the recommendations section of my assessment. Focus on quick wins."

The AI knows:

  • Your deliverable timeline (project description)
  • What you've already uncovered (memories from analysis chats)
  • The COO's skepticism of technology (memory)
  • Your preference for frameworks and bullet points (global context)

You get recommendations structured your way, sensitive to stakeholder dynamics.

Agency Owner

You: Digital marketing agency owner Global Context: Your agency focus, your need for client-ready outputs

Project: B2B SaaS Client Campaign

  • Description: Client goals, brand voice, competitive landscape
  • Files: Brand guidelines, previous campaign performance
  • Memories: What messaging resonated, what got rejected

A typical conversation:

You: "Generate 5 headline options for the new landing page."

The AI knows:

  • The target audience (project description)
  • The brand voice requirements (files)
  • That 'enterprise' language has been rejected before (memory)
  • You need client-presentable options (global context)

You get headlines that fit the brief and avoid past mistakes.

Multi-Passionate Individual

You: Explorer of multiple interests Global Context: Limited time, need to prioritize, quick learner

Project: Podcast Growth

  • Description: Current metrics, format, growth goals
  • Files: Episode analytics, audience survey results
  • Memories: What content performs best, distribution experiments

A typical conversation:

You: "I have 2 hours this week for podcast work. What should I prioritize?"

The AI knows:

  • Your growth goals (project description)
  • What's worked before (memories)
  • Your overall time constraints (global context)
  • Audience preferences (files)

You get a prioritized task list calibrated to your constraints and goals.

Context Compounds Over Time

The longer you use PromptBlocks, the smarter it gets:

Week 1: You set up global context and create a project Week 2: You've had several conversations; memories are building Month 1: The AI knows your preferences, past decisions, key constraints Month 3: Your context layer is rich — conversations feel like talking to a colleague who's been with you the whole time

You didn't maintain a document. You didn't update settings constantly. You just did your work, and your context built itself.

Tips for Maximizing Your Context Layer

Keep project descriptions focused and updated. When your goals change, update the description. The AI reads it every time.

Let memories accumulate naturally. Don't try to pre-load memories. Chat about your work and approve the useful suggestions.

Use prompts for repeated patterns. When you find yourself asking for the same thing in a similar way, save it as a prompt.

Create separate projects for separate contexts. Better to have multiple focused projects than one cluttered one.

Review memories periodically. Delete outdated ones. Quality over quantity.

What's Next

You now have a complete context layer:

  • Prompts for reusable instructions
  • Global Context for who you are
  • Projects for what you're working on
  • Files for reference materials
  • Memories for accumulated knowledge

From here, explore more:

You're ready to build your context layer.

The more you use PromptBlocks, the more it knows. Better context, better results. Your AI conversations just got smarter.