By Andreas Kviby, CTO, Stafe Development AB
For decades, software development has been bound by one simple metric: human time. Tasks were scoped, priced, and delivered based on how many developer hours were needed. But what if that entire foundation no longer made sense?
For decades, software development has been bound by one simple metric: human time. Tasks were scoped, priced, and delivered based on how many developer hours were needed. But what if that entire foundation no longer made sense?
At Stafe Development AB, we’ve spent the past four years building a new paradigm. A platform where time isn’t the bottleneck, AI becomes the engine, and developers focus on innovation, not repetition.
We call it 2DO.
What is 2DO?
What is 2DO?
2DO is not a project management tool.
It’s not a ticketing system.
It’s not a chatbot assistant.
It’s not a ticketing system.
It’s not a chatbot assistant.
2DO is an AI-driven task execution platform, directly connected to your codebase, environments, design systems, and infrastructure.
When you submit a TODO:
• 2DO analyzes the task description.
• It maps the request against your current application state.
• It identifies the affected code, files, or configurations.
• It autonomously proposes a solution.
• It implements the changes in a controlled environment.
• It generates screenshots, waits for your approval, and automatically deploys to production if validated.
Behind the scenes, a powerful AI engine running with 100+ GB RAM drives the reasoning, contextual understanding, code generation, and self-verification processes.
Humans are still part of the loop — planning, supervising, validating — but the heavy lifting is increasingly performed by autonomous AI agents trained on our internal knowledge and four years’ worth of real-world production code.
Why Build 2DO?
Why Build 2DO?
Our motivation was simple:
AI can now perform in minutes what a developer used to deliver in days.
If we kept selling human hours, we would be punishing customers for our own technological advancement.
Instead, we decided to build a system where customers pay for results — not hours.
Key goals:
• Automate the resolution of minor bugs, layout issues, and technical tasks.
• Empower human developers to focus on architecture, new features, and problem-solving.
• Ensure task consistency, faster cycle times, and immediate feedback loops.
• Enable scaling without growing operational overhead.
How Does 2DO Work in Practice?
Let me give you a real-world example:
On April 22, 2025, we had a backlog of small frontend issues:
• Button color mismatch
• Typo in service descriptions
• Incorrect mobile padding in admin panels
• Missing tooltips
• Minor UX label adjustments
Normally, these fixes would be estimated at 4 hours.
2DO handled the entire batch in 30 minutes:
• Automatic issue parsing
• Codebase scanning
• Patch proposal generation
• Staging deployment
• Screenshot validation
• Production deployment via GitHub
This is not theory. This is happening now.
Under the Hood: How 2DO is Built
2DO combines several components into one seamless experience:
• Windsurf AI Code Editor is used to securely and efficiently write, validate, and commit code changes triggered by AI agents.
• Markdown file hacks are employed to structure, tag, and manage TODO tasks in a lightweight, version-controlled manner that AI agents can easily interpret.
• GitHub serves as the backbone for version control, pull requests, branch handling, and deployment.
• A custom-built Filament PHP Todo system acts as the operational interface where tasks are created, updated, prioritized, and monitored — fully integrated with GitHub to sync task statuses automatically based on branch merges and commits.
All components are orchestrated to maintain a safe, auditable, and transparent flow from idea → task → execution → production deployment.
How 2DO Changes the Customer Relationship
When clients sign up for 2DO, they aren’t just buying access to developers.
They’re buying an AI-powered production engine, supervised by senior engineers.
Our model offers:
• 16 scheduled human meeting hours per month (workshops, planning, consulting)
• Unlimited access to 2DO’s AI task execution for everyday operational tasks
• Dashboards showing AI utilization vs. traditional human working hours
In many cases, clients can expect up to 200 hours’ worth of human-equivalent work per month — for a fixed, predictable investment.
Meanwhile, our human engineers are freed to focus on the most valuable work: scaling platforms, optimizing performance, launching new features.
Final Thoughts
2DO is not about replacing humans. It’s about amplifying human potential.
It’s about freeing developers from tedious maintenance work.
It’s about giving customers faster results, higher ROI, and more control over their platforms.
At Stafe Development AB, we believe this is the future of software development. And the future isn’t coming someday — it’s here now.
If you’re ready to work smarter, not harder —
If you’re ready to scale without growing headaches —
If you’re ready to let AI and humans collaborate seamlessly —
Then 2DO is ready for you to try when we launch it to the waiting-list. Contact we if you are interested in testing it in September 2025. Then we will launch it in public beta.
Andreas Kviby
CTO, Stafe Development AB
📧 andreas@stafegroup.com
Andreas Kviby
CTO, Stafe Development AB
📧 andreas@stafegroup.com