fatih kadir akın

@fka.dev

happy human • ⭐️ @github.com star • devrel @automattic.com for @wordpress.com • windsurf ambassador • developer • board member x.com/avtedorg • writing books • entrepreneμr (2 exits) • fka.dev • github.com/f • prompts.chat • x.com/fkadev

✨ my AI-based git history rewriter hit 710 stars in just 5 days 🤩 Absolutely wild! I love that all the commit-history OCD folks are actually using it 😂 v0.6.9 now supports Windows and includes security improvements like diff redaction 🔒 👉 github.com/f/git-rewrit...

Bild

🔥 Last week I ran a workshop at Amazon Türkiye office, diving deep into Amazon’s brand-new agentic IDE kiro.dev, exploring what it can really do and how it stacks up against other agentic code editors. Over 100 people joined and the energy was incredible! ✨

BildBild

After 10+ years of open-source and community contributions, seeing my name on a star on the floor, in the heart of tech, at one of the best developer conferences, feels like being a Hollywood star 🥹🤩 I love @github.com

Bild

I made a simple chart to show the core differences between LLM, RAG, and MCP. It illustrates how each layer adds new abilities, from plain model reasoning to retrieval, real-time actions, and finally, standardized tool calling through Model Context Protocol.

Bild

My October event calendar: 🗓️ Oct 10 – SistersLab Git & GitHub 📍 Online 🗓️ Oct 11 – SaaS Summit 2025 📍 Şişli 🗓️ Oct 18 – Komünite MCP Bootcamp 📍 Vadi İstanbul 🗓️ Oct 22 – AI Summit 2025 📍 Ankara, Ministry of Youth & Sports 🗓️ Oct 24-31 – GitHub Universe 📍 San Francisco

🏄‍♂️ I'll be trying this prompt to refactor some blindly vibe-coded codebase. Let’s see how it goes. Any extra recommendations? (Plain prompt text is in ALT)

Refactor with this structured flow:
1. Identify bloated files: Detect files that are unusually long or overloaded with logic.
2. Analyze code organization: Understand how modules, functions, and dependencies are structured to spot opportunities for cleaner separation.
3. Modularize components: Break down large files into smaller, reusable, and well-scoped components, functions, or utilities.
4. Eliminate duplication: Consolidate repeated code into shared modules or helpers to reduce redundancy and improve maintainability.
5. Ensure clarity & reusability: After refactoring, verify that the new structure is easier to navigate, extend, and test.

🔥 I just shipped some exciting updates to my prompt embedding app, prompts.chat/embed! It now supports file trees and diff views, perfect for making your posts and app docs more interactive and clear. And of course, it’s open source! (Completely vibecoded by Claude 4 Opus & Sonnet)

🚀 New Post: MCP isn't just about tools! It has progress tracking, cancellation, real-time subscriptions, AI sampling, and more advanced features that make AI assistants feel interactive. Wrote a comprehensive guide with examples👇 blog.fka.dev/blog/2025-06...

Diving into MCP Advanced Server Capabilities: A Comprehensive Guide

An in-depth exploration of advanced Model Context Protocol features including progress tracking, cancellation, completions, logging, and more with practical implementation examples.

blog.fka.dev

🔥 NEW POST: Last month, MCP made a big change: They moved from SSE to Streamable HTTP for remote servers. It’s actually a pretty smart upgrade. If you’re building MCP servers, this change makes your life easier. I've explained why: blog.fka.dev/blog/2025-06...

Why MCP Deprecated SSE and Went with Streamable HTTP

An exploration of why the Model Context Protocol is moving from Server-Sent Events (SSE) to Streamable HTTP transport, the benefits of this change, and what developers need to know for future MCP impl...

blog.fka.dev

MCP is designed as programmatically callable interfaces using HTTP-compatible JSON-RPC. So instead of building separate RESTful endpoints, why not expose your functionality through only the MCP server? It simplifies integration with LLMs and keeps your API surface consistent and unified.

MCP has announced "elicitations" as part of the protocol (as draft) which made me excited! 🔥 Just wrote about how it standardizes interactive AI workflows - basically formalizing the AI-generated UI concept I was exploring 👇 blog.fka.dev/blog/2025-01...

MCP Elicitations: Standardizing Interactive AI Workflows Through Structured User Input

Explore how the new MCP Elicitations protocol standardizes interactive AI workflows by enabling servers to collect structured user input dynamically, building upon AI-generated UI concepts.

blog.fka.dev

17 yıllık kariyer yolculuğumda öğrendiğim bazı bireysel şeyler (tavsiye değil, şahsi çıkarımlarım): - Başlamayı fazla düşünmemeliyim, direkt başlamalıyım. Devamı yolda şekillenir.