This is a really well-thought-out implementation. I like that you let users bring their own OpenAI API key—it removes a major barrier for creators who want to test the app without you covering inference costs.
The workflow from topic generation → post creation → hashtags → optional image feels practical rather than just being another AI demo. I also appreciate that you’ve shared both the n8n workflow and the WeWeb template, which makes it much easier for others to learn from and extend.
One feature I’d personally love to see is a brand voice memory. Instead of only relying on a prompt, users could upload a few of their best-performing LinkedIn posts, and the AI could learn their writing style, preferred structure, and tone over time. Combined with feedback on post performance, that could make the content feel much more personal and less generic.
It would also be interesting to add support for multiple LLM providers (OpenAI, Anthropic, Google Gemini, etc.) through a unified interface, giving users more flexibility based on cost and output quality.
Overall, this is a solid foundation for anyone looking to build an AI SaaS around LinkedIn content creation. Thanks for sharing the templates and tutorial—I think they’ll be valuable for both beginners and experienced builders.
