I got tired of rewriting the same content for 9 different platforms. So I built Repostify.
For the past couple of years, I’ve been creating long-form content and then manually adapting it for different platforms — LinkedIn, X, Telegram, newsletters, sometimes Instagram and YouTube Shorts.
Every single time it was the same frustrating process: copy the text → rewrite it → adjust the tone → shorten it → reformat. It took way too much time and mental energy.
A few months ago I decided to solve this problem for myself. I started building Repostify — a tool that takes one piece of content and turns it into multiple platform-ready versions while trying to keep your own voice and style.
The main idea is pretty simple: instead of generic AI rewriting, it learns your style and applies it across different platforms.
Right now the full experience lives in a Telegram Mini App (that’s where I can ship updates the fastest). You paste a long text and get ready-to-post versions for LinkedIn, X, Telegram, Email and other platforms. The web version is still in development.
It’s completely free to start — no credit card, no complicated setup.
I’m building this in public and would genuinely appreciate your feedback:
How do you currently deal with repurposing content across platforms?
Would a tool like this be useful for you?
What’s missing or could be improved?
If you want to try it, just open Telegram and search for Repostify, or use this link:
https://t.me/repostifyai_bot
Happy to share more details about how it’s built or discuss the challenges I’m facing right now.
Also happy to share the tech stack if anyone’s curious (Next.js + Supabase + Grok).
The cross-platform repurposing problem is one of those pains that's obvious in hindsight. Every content team we've talked to has either built a messy internal tool for this or lives in a spreadsheet. The tricky part, technically, is that "same content, different platform" is deceptive; LinkedIn and Twitter don't just need different lengths; they need different narrative structures (LinkedIn = story arc, Twitter = hot take first). Curious whether Repostify is doing rule-based transformation, LLM-based restructuring, or both. The teams that get the most out of these tools are the ones where the AI preserves the author's specific voice, not just reformats the text. How are you handling tone/voice consistency across transformations?
the core tension in content repurposing tools is that the platforms that need the most adaptation, LinkedIn to X being the classic example, also require the most judgment about what to cut and what to emphasize. a LinkedIn post builds to a conclusion. an X thread starts with the hook. those aren't the same content adapted, they're different editorial decisions. how does Repostify handle that structural difference and is it making those decisions or asking you to make them
You're right — that's not a formatting problem, it's an editorial one. LinkedIn builds to a conclusion; X leads with the hook. Same ideas, different narrative architecture.
Repostify treats them as separate drafts, not resized copies. The model gets explicit per-platform constraints: LinkedIn gets a bold first-line hook + 3–5 paragraphs that build; X gets a punchy opener, numbered thread pacing, and a strong CTA on the last tweet. Same source material, different structure — not "LinkedIn post, but shorter."
On editorial judgment: the AI makes the first pass — what to lead with, what to cut, what to emphasize per channel. You review all 9 versions side by side and edit before posting. We clone your voice from 3+ real posts (tone, emoji density, CTA style), but the structural decisions are prompt-driven defaults, not a wizard asking "hook-first or conclusion-first?"
Honest take: if you want full control over every editorial fork, you'll still edit. Repostify is a strong first draft that respects platform grammar, not a autopilot that replaces your judgment.
Smart angle. I do the same for my own content — write once, syndicate to X/IndieHackers. The Grok idea is clever because it forces you to optimize for tone per platform, not just length. How do you handle the longer-form posts that get cut off by the preview?
Same workflow here — write once, distribute everywhere. The preview fold is exactly why we don't treat syndication as "same post, shorter."
For feed surfaces (LinkedIn, Instagram), the prompt forces a standalone hook above the fold — bold first line on LinkedIn, ≤2 visible lines on IG — then the body earns the "see more" click. For X, it's a thread from the start, so nothing important hides behind a single-post truncation.
The longer narrative doesn't get squeezed into the preview. It goes to Medium (400–700 words) or email (200–350 words). Social posts are entry points, not the full essay pasted into a feed card.
No preview simulator in the app yet — you eyeball the hook before posting. If the fold line feels weak, that's the one edit worth making.
You asked what could be improved, so I looked at the site from the technical side. Speed is already good, so I focused on where Lighthouse actually flags things: accessibility (scores 79) and SEO (85). Most of it is quick to fix, and several items also help your visibility, which matters a lot for a tool like this.
In rough priority order:
Readability. A fair bit of the text is low contrast and small. Blue text sits on pale blue backgrounds, and some grey text on white is light, much of it around 10 to 13px. Since this page's main job right now is to convince someone to open the Telegram bot, the words doing that selling should be crisp and easy to read. Darkening the blue and the grey a few steps fixes most of it.
Your images have no alt text, and your form fields have no labels. Two birds here. Alt text helps screen readers and helps you show up in image search, and labels make the form usable for everyone. For a content tool, getting these basics right is also just on brand.
Your headings skip levels, for example jumping from an H1 straight to an H3. Putting them in order is a small change that helps both screen readers and search engines understand the page.
Your robots.txt is flagged as invalid. For someone who wants to be found, that's worth a quick fix so search crawlers read your rules correctly.
None of this is urgent, and the page is fast (performance is 96, best practices a perfect 100), so the bones are good. But since you're building in public and you lean on being discovered, cleaning up the accessibility and SEO basics is probably the highest-value hour you could spend on the site right now.
One bigger-picture note. Your real product lives in Telegram, and the site is mostly a doorway to it. So the single most important thing this page does is get someone to tap through. Everything above serves that: make the pitch easy to read, and make the page easy to find.
Thanks Johan, really appreciate you taking the time to review the site and share the Lighthouse findings.
You're right — the core experience is already fast and clean. We'll work on the accessibility and SEO improvements you mentioned (contrast, labels, headings, and robots.txt).
Thanks again for the thoughtful feedback!
I build in this exact space, so this one is close to home. The thing most repurposing tools get wrong: they treat it as shorten plus reformat, when platforms actually differ in what kind of content wins, not just length. LinkedIn rewards a story or a stance. X rewards one sharp idea stripped of everything else. A YouTube Short dies if the hook is not in the first two seconds. A tool that reformats the same angle five ways produces five mediocre posts. The version that performs reframes the angle per platform. Two things I would protect as your moat: voice fidelity (if it reads like generic AI, nobody ships it under their name) and real per-platform reframing, not just resizing. Nail those and you are not competing with the paste-into-ChatGPT crowd anymore. What does retention look like so far, are people coming back weekly or using it once and drifting?
Building in the same space means you already know the failure mode — five resized versions of the same angle. We designed around that from day one.
Voice fidelity: before any generation, you paste 3+ real posts you've actually shipped. Grok mirrors your tone, vocabulary, emoji density, and CTA patterns — not a generic "creator voice." If it reads like AI, you don't post it. That's the bar.
Per-platform reframing: each surface gets its own editorial contract, not a character limit. LinkedIn → story/stance, 3–5 paragraphs building to insight. X → one sharp idea per tweet, threaded from a punchy opener. YouTube Shorts/TikTok → hook in the first 2–3 seconds with timed markers. Medium/email get the full narrative arc. Same source material, different angles — the prompt example │ literally shows LinkedIn building to a conclusion while the X thread leads with the hook and distributes lessons across tweets.
On retention: still early — Telegram-first launch, 120+ creators active there, Product Hunt is the web push. What we're seeing so far matches publishing cadence: creators who ship weekly come back weekly (new article → new batch of 9 formats). The one-and-done "paste once, never return" cohort exists too — usually people who generated but didn't actually post anything. Signals we watch: repeat generations, history revisits, whether they saved "My Style." retention is the metric that matters for us, not curiosity clicks. Too early to quote clean cohort numbers publicly, but the pattern is encouraging for the weekly publishers — that's who we built for.
Would love to compare notes offline if you're open to it — always curious how others are solving the reframing layer.
The "learn your style and apply it across platforms" framing is the right differentiator. Generic AI rewriting exists everywhere. Voice-consistent repurposing is the harder and more valuable problem.
One thing to pressure-test: platform format differences go deeper than tone. LinkedIn rewards long paragraphs and story arcs. X punishes them. Telegram channels have different reading patterns than newsletter subscribers. "Same voice, different format" is the goal, but format is doing a lot of work there — it's not just length.
The Telegram Mini App as the primary surface is a smart distribution decision while the web version is being built. It removes signup friction entirely for users who are already in Telegram. Worth watching whether your early users are using it primarily in Telegram or asking for a web version — that answer tells you a lot about who the tool actually fits.
What does "learning your style" look like in practice right now — are you analyzing samples the user provides, or is it more about the platform-specific output rules?
Format does the heavy lifting — agreed. We split it: hardcoded per-platform editorial rules (LinkedIn story arc, X threaded hooks, Telegram skim-format, email subject-first) + in-context voice cloning from 3+ real posts you provide each run. No persistent style profile yet — samples go raw into the prompt, Grok mirrors tone/emoji/CTA across all 9 surfaces.
Telegram-first was intentional (zero friction). Early users are mostly TMA; PH is testing web demand via waitlist. That split will define who we build for.
that's good
What makes this tricky is that the same behavior can support very different interpretations.
Someone repurposing content across platforms could be solving a distribution problem, a consistency problem, a workflow problem, or something else entirely.
That's why I'd be hesitant to treat "rewriting content" as the validated pain too early.
The interesting part is often the decision that gets made from the observation, not the observation itself.
Thanks for the sharp observation — you're right, and it's something I've been thinking about a lot.
"Repurposing content" is indeed a behavior, not a single pain point. From what I've seen so far, people do it for different reasons:
Some are primarily fighting time (they have good long-form content but hate manually rewriting it 5–8 times).
Some care more about consistency of voice and messaging across platforms.
Others are focused on distribution — they know they should be posting more widely but the friction stops them.
A few are trying to solve creative fatigue (they run out of ideas for how to adapt the same piece).
The current MVP (Telegram Mini App) is mostly aimed at the first group — people who want to go from one piece of content to multiple platform-ready versions quickly, while still sounding like themselves. That's the clearest signal I've gotten from early users.
That said, you're making a good point: just because people do repurpose content doesn't automatically mean it's a painful enough problem for them to pay for a dedicated tool. I'm still in the phase of trying to understand which interpretation of the problem is strongest (and for whom).
Would you be open to sharing how you usually distinguish between surface behavior and the actual underlying job-to-be-done when validating ideas?
What's interesting is that you've already identified several plausible interpretations.
The hard part is figuring out which one actually deserves confidence before the product starts optimizing around it.
That's not something I'd try to answer properly in a thread.
If you're curious, drop your email and I'll send over the tighter version.
Thanks, I'd appreciate reading the tighter version.
You can reach me at: [email protected]
Looking forward to it.
Sent you a note by email.
I think the interpretation decision matters more than the repurposing behavior itself right now.
I can relate to this.
Before building AmpPilot - AI marketing platform, I was doing something very similar — a mix of GPT-5, Claude, Canva, and a scheduling tool like Buffer. Multiple permutations, multiple prompts, and still a lot of manual work stitching everything together.
What bothered me wasn't just rewriting content for different platforms. It was that every post needed context — brand voice, what we'd already published, what competitors were saying, what was working historically, and the constant need to decide what to post next.
That's actually what pushed me to build AmpPilot. Instead of generating one post and manually adapting it, I wanted the system to continuously learn the brand context as it evolves and handle the entire workflow end-to-end. Users simply connect their channels, and content gets created and published automatically.
Interesting to see how many founders arrive at similar problems from different angles.
Thanks for sharing your experience — I can relate to that pain too.
You're right that the real bottleneck often goes beyond just rewriting. Context, brand voice consistency over time, historical performance, and deciding what to post next are all significant challenges.
What we're building with Repostify is a bit more focused: helping creators quickly turn one strong piece of content into high-quality, platform-native versions while preserving their own voice. A lot of the founders and creators we’ve spoken with still prefer having control over individual pieces rather than going fully autonomous.
We see it as solving the “I have this long post/article/thread and need to distribute it properly across platforms without losing quality or sounding generic” problem. It’s more of a high-signal repurposing tool than a fully autonomous content engine.
Interesting how different teams end up attacking adjacent parts of the same big problem.
Would love to hear more about how AmpPilot handles brand context over time — sounds like a really hard (but valuable) thing to get right.
Agreed. As you mentioned and I have heard many founders say that they do not want things to be done with one click or fully autonomous. That's dangerous too.
At AmpPilot, though we have AI handling every post, blog design automatically, but it also gives you an option to review things and then schedule it or take it live or make changes if needed, be it design or content. We call this as "Ask-first mode"
Plus point is you can switch this to autonomous to "Ask-first mode" for any stage any time as well.
This comment was deleted a day ago.