Been building react.js & next.js apps for a while and rich text editors are honestly one of the most painful pieces to integrate as I've already tried tinymce, tiptap etc and getting AI features wired in cleanly is always a separate battle.
the part that always slows me down is keeping editor logic and AI logic in sync - prompt updates, error handling, streaming states all end up in two different layers and get messy fast, having both bundled into one SDK is a huge time save if it works the way you describe.
lexical-based + drop-in npm + BYO keys is exactly the right approach! definitely trying this on a client project next week 👏
Hey Product Hunt — Nimish here, founder of Eddyter.
Eddyter is an AI-powered rich text editor built on Lexical that you install via npm and drop into any React or Next.js app. Think of it as the editor component you'd build yourself if you had six months — with AI features (generate, refine, translate, summarise) already wired in.
What makes it different: • Editor-first, not chat-first — AI lives inside the writing flow • BYO API keys (OpenAI, Anthropic) or use our managed credits • 30-minute integration, works with Cursor / Claude / Lovable
🔗 Docs & quickstart: https://eddyter.com/docs
📦 npm: https://www.npmjs.com/package/eddyter
🎥 What is Eddyter? https://youtu.be/oNHBa-DImZc
🎥 Integrate in 10 min with AI tools: https://youtu.be/5lTjRFjUWgs
Would love feedback from devs shipping AI features — what editor pain points are you hitting?