Zack Proser

For me, queued messages are the biggest DevEx unlock for @cursor_ai.

Now I can enter a flow state where I'm three feature requests/bug fixes ahead of Cursor.

I rapidly verify functionality and put new tasks on the queue while it's working through my requests ❤️

workos.com/blog/mcpnig...

Interest in, and adoption of, MCP is exploding.

Think: one spec → any tool, any workflow, securely. Big shift for AI infra.

Recap from MCPNight here: #MCP #AIInfra

Learning your stack deeply isn’t just about raw velocity—it’s about agility.

When requirements shift, you can change direction gracefully, leveraging your expertise rather than scrambling for docs.

Deep expertise isn’t about being a "rockstar"—it’s about sustained rapid iteration.

The tighter the feedback loop, the faster you evolve.

This combination of open-source + tutorials has been incredibly rewarding.

There’s nothing better than hearing, “Oh, never mind, I just found your tutorial.”

What’s working well for you in Developer Education? Let’s compare notes!

What makes this approach effective:

• Hands-on learning: Developers can clone, run, and modify the repos.

• Immediate value: Tutorials bridge the gap between code and implementation.

• Real-world examples: These aren’t toy projects—they solve actual problems.

Document Access Control with AWS CDK + Lambda Authorizers

☁️ Repo: Full serverless stack (API Gateway + S3 + Lambda).

🚀 Tutorial: workos.com/blog/how-to...

✅ Securing RAG Applications with Fine-Grained Authorization

📂 Repo: Pinecone vector DB + WorkOS FGA for document access control.

🔐 Tutorial: workos.com/blog/how-to...

Here are some recent examples:

✅ Browser-Based OAuth for CLI Tools

🖥️ Repo: Secure token fetching & storage for CLI apps.

📚 Tutorial: workos.com/blog/how-to...

The formula:

1. Build an example architecture for a real-world use case.

2. Open-source the code as a companion repository.

3. Write a tutorial that explains how to implement it step by step.

This combo makes concepts actionable.

One thing that’s working well for Developer Education at WorkOS: pairing open-source companion repositories with detailed tutorials.

It’s helping devs learn faster, solve complex problems, and build better systems. Here’s how we do it: 🧵

Thanks so much 🙏😀

Interactive experience here: zackproser.com/demos/tokenize

Hey, I’m Zack 👋

13-yr full-stack dev #buildinginpublic to help others learn.

Today’s toy: an interactive token-izer—paste any text, watch an LLM slice it up (👀↓).

Into no-fluff deep dives on AI/ML, RAG & #Next.js? Hit Follow and let’s geek out.

Getting deep with a single stack means fewer overlooked edge cases.

You see pitfalls coming from a mile away.

Production issues decrease, and deployment confidence soars.

Sticking to a known stack, you internalize common patterns and best practices.

That means fewer surprises, simpler refactors, and faster scaling when your user base grows.

Do you tend to use multiple stacks or have one or two favorites?

Shiny features are tempting.

But in a micro #SaaS, your job is to validate, refine, and ship quickly.

Ask your initial customers what’s missing and fix that first.

Deep expertise fosters stronger ties to your stack’s community.

That means direct access to new tools, timely support, and a constant influx of fresh ideas—further feeding your speed loop.

Micro SaaS can be a small side project—or the seed of your next major venture.

Keep your overhead light, set achievable milestones, and watch your momentum grow.

Try It Out

Customize your YAML, define your own agent roles, or pick a new game concept—CrewAI will coordinate the entire AI-driven pipeline.

Why CrewAI Matters

It’s not just for games.

You can chain specialized agents for anything: blog posts (writer + editor), product specs (engineering + legal), or doc creation (researcher + fact-checker).

Get It Running

Create a new Poetry project, install dependencies, and set your OpenAI key.

poetry run python3 main.py will prompt you to choose a game scenario and produce a fully functional Python script.

It’s all run sequentially with a Crew:

1. Code generation by the Senior Engineer agent
2. QA checks by the QA Engineer
3. Final sign-off from the Chief QA Engineer

How to build a game-building agent system with CrewAI — WorkOS A hands-on guide with patterns, agents, and executable examples

Example Setup

Agents YAML: Defines your game dev, QA, and chief QA roles.

Tasks YAML: Step-by-step instructions (generate code → review → final validation).

Game Design YAML: The actual game requirements (Pac-Man style or whichever idea you fancy).

workos.com/blog/how-to...

Core Concepts

Crews: Groups of specialized agents (e.g., a “Senior Engineer” generating code, QA agent reviewing).

Flows: Event-driven workflows, giving you fine-grained control over who does what, when.

AI Agents That Code Your Game—Meet CrewAI

What is it?

CrewAI is a Python framework for orchestrating multiple AI agents (like a mini dev team) to generate, review, and validate code. Perfect for building something fun—like a Pac-Man or Pong clone—in Python.

Pick a specific problem and solve it exceptionally well.

Micro #SaaS thrives on niche focus.

It’s better to delight a small audience than to be “okay” for everyone.

Sure, it’s a new toolkit.

But once you “get” streaming and how the same interface adapts across multiple LLMs, you’ll be glad you made the investment.

RAG Example: I built a “chat-with-my-data” tool using the AI SDK + Pinecone for retrieval. The pipeline: user question → embed + search → context injection → stream answer + show relevant docs.