GoodSocials
AI social media manager for LinkedIn. Only authentic content
156 followers
AI social media manager for LinkedIn. Only authentic content
156 followers
AI social media manager for LinkedIn. Writes high quality content, you approve it in a kanban board, it schedules so you post regularly, close to daily.
Payment Required
Launch Team / Built With
Framer AI AgentsDesign and publish professional sites with AI
Promoted
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
@pavelk2 I run a similar pipeline for my own project, summarising merged PRs into posts, and it bit me last week: it announced a feature as live because the PR shipped the terms for it, while the switch that actually turns it on lived in a different file. How do you stop it from announcing work that's merged but not yet released? Do you read feature flags or deploy state, or rely on the kanban review to catch it?
@pavelk2 Congrats on the launch !
Does it only work with LinkedIn? Are you planning to expand and support more social media platforms?
@adana Thanks!
Currently goodsocials only supports LinkedIn. Obviously as the name suggests, other socials are coming. Instagram, and X will go next.
@pavelk2 that's good !
Good luck =)
Happy to see a new platform for LinkedIn. :)
What made you to choose this one? :)
@busmark_w_nika THanks Nika!
As I wrote in my intro comment - I was posting regularly on LinkedIn for TimeTuna, and was looking to find a way to simplify the process. I am not heavy on X, but I guess that should be the next step.
What would you recommend?
@pavelk2 I am also thinking whether more companies didn't try Reddit marketing too.
@busmark_w_nika good one!
The LinkedIn ghosting era might finally be over haha. You still get to say no in the kanban, which is the important part. Nice work! @pavelk2
@zach_francis THanks Zach!
The approval-first workflow is thoughtful—how do you measure whether scheduled posts outperform manual drafts?
@michael_astreiko I don't expect scheduled posts to outperform manual ones. We focus on consistency of posting, and helping to publish more, and publish often.
"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
@galdayan this is hard. We did not crack it fully yet. But I think we will. I introduced 10 levels of account. At level 1 - the tool does not know much your preferences. At level 10 - it should kinda speak in a way that you feel very comfortable sharing such content, and your followers appreciate what it writes.
@pavelk2 10 levels is a smarter framing than I expected - most tools just have one global tone slider. honest answer on "not cracked it fully" is worth more than a confident claim would've been. good luck with it, genuinely curious to see where level 10 lands for people once you've got enough usage data to tune it.