GoodSocials

"only authentic content" is the hard part to prove, not the scheduling. we post on PH and LinkedIn for Dial and half the battle is just not sounding like every other AI post out there, generic pro/con structure, semicolons everywhere. how does the kanban review step actually catch that before it goes out, or is it still on the human approving to notice and rewrite it

The approval-first workflow is thoughtful—how do you measure whether scheduled posts outperform manual drafts?

Happy to see a new platform for LinkedIn. :)What made you to choose this one? :)

@pavelk2 Congrats on the launch !Does it only work with LinkedIn? Are you planning to expand and support more social media platforms?

Hi ProductHunt 🙂 I previously built @TimeTuna - gorgeous scheduling pages (#1 on producthunt in Dec 18 2025). I love the idea of weekly updates on linkedin about the product I work on. I used to write them manually: what I built, and what I learnt. I took me around 1 hour to write such post. It felt long, and I did not want to use AI for improving it, as I hate AI slop.But I knew that AI is great at summarising information. So I asked it to summarise what I did last week by reading my pull reqeusts from Github, and meetings in my google calendar. I asked it to summarise interesting findings from the behavior of users on PostHog, Stripe, and some logs and bugs from Vercel and Stripe. In addiion I asked it to do a deep research on what happened in the fields of scheduling meetings, and meeting software (think Calendly, and Cal.com), and write down some insights from the market research. I used Claude Code to combine everything in one place. I was very pleased with the content, but not the tone of voice. I wanted it to speak as a researcher rather than a self-promotion sloppy machine. After 10 iterations it became soemthing I am happy with. Then I added some filters on what to never expose (e.g. my personal data, personal data of people I met with, my customers etc, not to expose absolute numbers but speak about trends, and percentage changes). Then I connected Canva and Flora AI to generate an awesome visual as well.Few people say those posts are great and they appreciate that I write them. I don't write them. Every thursday evening a routine on Claude Code runs a system of skills connected to 8 different MCPs does the job and schedules a post via Buffer. I have time till Friday morning to check it out, and correct if I want. I realised that I would love to help other people create such relevant interesting posts, not based on slop, but based on data (either your own integrations, or based on deep research). I don't mind AI content - if it is high quality (e.g. think - about top 10 startup incubators in the UK - I would probably read it and trust it more if it is written by a human).Anyway. that is a story behind @GoodSocials . I would love your feedback. Peace. Love. No AI cringe.P.S. Get a 33% discount if you want: #NOCRINGE33