A working journal from the studio. Essays on AI, agents, retrieval, and the considered work in between, written by the people doing it.
Latest essay · 4 min
Turkey runs its entire justice system on one closed-source, government-mandated editor — and on Apple Silicon it was barely usable. I rebuilt it from the outside with Claude Code as a debugging partner: twenty fixes, no source code, no vendor cooperation.
Said SürücüShowing 5 of 12
3 essays
4 min readVideo
The first version was a progressive web app. But I decided the best way to fix certain things was to chart the right course from the very beginning.
Onat VuralClaude Agent SDK, Software Development, AI, Indie Hacking7 min read
After our 587-star Yargı MCP, why did we build a CLI too? Because the real language AI agents use to talk to the world isn't protocols — it's the command line.
Said SürücüCLI, AI, MCP, AI Agents, Legal Tech7 min read
If your job involves staring at a screen for 8 hours a day, there's a good chance you won't be doing it two years from now.
Onat VuralAI, Future of Work, Business, Agentic AI2 essays
11 min read
An autonomous financial analysis platform built on multi-agent AI architecture designed for Turkish financial markets. Achieving 50-70% performance improvement through parallel task execution.
Said SürücüAI, Multi-Agent, Finance, Architecture6 min read
The fundamental challenge preventing AI from becoming truly useful in our daily lives—and how we solved it.
Onat VuralAI, UX, Memory, ArchitectureSaid Sürücü
Onat Vural
Beta Space Studio Team
