This is a submission for the DEV Weekend Challenge: Community.
Why I Built This (And How)
I didn’t start this project with a prompt.
I started with a problem.
I write technical posts regularly. I genuinely enjoy the writing part.
What I don’t enjoy is rewriting the same idea four different ways for four different platforms.
DEV → X → LinkedIn → Reddit → Substack.
Same thinking. Different tone. Different format. Different constraints.
That friction slowly turns “sharing your work” into overhead.
So before I wrote a single line of code, I wrote:
- a short spec
- a high-level PRD
- clear scope boundaries (what ships vs what doesn’t)
Only then did I bring in AI.
ContentSeed was built for a community I’m part of every day: developer-writers — people who want to publish consistently without spending more time repackaging than creating.
This project is fully solo.
And intentionally disciplined.
The goal wasn’t to build “an AI app.”
The goal was to build a small, focused tool that:
- respects privacy
- stays client-side
- lets you switch between LLMs
- solves one real pain point well
What I Built
ContentSeed is a client-side content repurposing app that turns one technical post into platform-native drafts for:
- X / Twitter (thread-style)
- Substack
Demo
YouTube demo:
Live app:
https://contentseed.vercel.app
GitHub:
https://github.com/mohitSharma74/contentseed
Core Principles
1) Easy and Accessible
Paste content → pick platform → generate → tweak → export.
No accounts.
No dashboards.
No friction.
2) Privacy-First
There is no backend.
No telemetry.
If you bring your own API key, it stays local.
3) Model Freedom
You’re not locked into one provider. You can generate using:
- Minimax
- OpenAI
- Anthropic
- Gemini
Use what you already pay for.
AI as Collaborator, Not Crutch
I care deeply about how AI is used.
For this project, I didn’t “vibe code.”
I wrote the spec and high-level PRD first.
Then:
- I used Minimax M2.5 for the majority of implementation.
- I used gpt-5.3-codex (high reasoning) for:
- validation
- edge-case analysis
- testing logic
- architectural review
I intentionally separated:
- generation from validation
- speed from correctness
- features from product clarity
That separation mattered.
It kept the app:
- fully client-side
- privacy-first
- small in scope
- focused
There were features I wanted badly — image generation, short-form video generation — but they didn’t make the cut.
Not because they weren’t exciting.
Because they weren’t essential.
What It Can Do Today
- Platform-native drafts (not copy/paste spam)
- Demo mode (no API key required)
- Tone controls: Casual / Professional / Technical / Storytelling
- Length controls: Shorter / Default / Longer
- Toggle hashtags and emojis
- Copy / regenerate per platform
- Export as PNG, JPG, or PDF
A Small Tool. On Purpose.
ContentSeed is not trying to become a content platform.
It’s not a dashboard.
It’s not a growth engine.
It’s not a subscription funnel.
It’s a utility.
It exists to remove one specific kind of friction:
rewriting the same idea for different platforms.
It stays client-side.
It stays privacy-first.
It stays model-agnostic.
If it grows, it will grow deliberately.
For now, it does one thing well.
And that’s enough.
Top comments (0)