Open by default
We document decisions, learnings and trade-offs so others can reuse the path, not just inspect the outcome.
A lab where ideas become real products: agentic AI, AI tools, developer-first infrastructure, learning apps and open source GitHub experiments anyone can learn from, contribute to or build on top of.
Ideas shared from day zero
Prototypes with real code
Public repos and traceability
We document decisions, learnings and trade-offs so others can reuse the path, not just inspect the outcome.
We build agents, RAG systems and automations that solve operational work, not empty demos for hype.
Design, architecture, DX, QA and observability matter from prototype stage so an experiment can grow.
The door is open for builders, designers, founders, students and teams who want to learn by shipping. Our strong focus: applied agentic AI, reliable software and real product work.
Useful agents for real work: research, support, operations, automation and education.
Tools to create, validate, deploy and observe software with less friction.
Educational experiments that combine narrative, guided practice and immediate feedback.
Small products that solve everyday problems for communities, businesses and local teams.
These repos are part of Wasyra's public GitHub: prototypes, agents, RAG systems, design systems and tools where we experiment with agentic AI and productive software.
github.com/wasyra
Experimental agent for reservation flows, operational coordination and conversational automation.
github.com/wasyra
Atomic React design system: Tailwind v4, Radix, Motion, OKLCH tokens and MCP-friendly catalog.
github.com/wasyra
WhatsApp agent experiment for business operations with automated conversational workflows.
github.com/wasyra
Agentic RAG lab for search, reasoning and LLM-powered knowledge systems.
github.com/WasyraTech
Multi-agent editorial room with CrewAI: Tavily research, draft and LinkedIn post editing in a sequential pipeline.
github.com/WasyraTech
Same editorial use case with LangGraph: typed state, explicit graph and editor↔writer loop until approved.
We define the problem, user, minimum scope and why it deserves to be built in the open.
Design, architecture and the first functional build with recorded decisions and visible debt.
Demos, feedback, issues, contributions and the next version based on real usage.
Bring the problem. We bring agentic AI, design, architecture, product thinking and a clear way to build it in public.