Show IH: A portable voice profile for your AI writing and agents
Hey Indie Hackers,
I’m building Noren, a tool that turns writing samples into a portable AI voice profile.
The problem I’m trying to solve: AI writing tools can produce clean drafts, but they often erase the way you actually write. Most tools handle this with tone settings like “friendly” or “professional.”
Noren instead extracts patterns from real writing samples: sentence rhythm, structure, phrasing, analogy habits, things you tend to avoid and a lot more.
So far, the clearest early users have been romance writers and newsletter writers. Romance writers care about preserving character/narrative voice. Newsletter writers care about making AI useful without turning every issue into generic content-marketing prose.
I also think this becomes more important as people use agents. If an agent is replying, drafting, summarizing, or negotiating on your behalf, it probably should not sound like the model default. It should have access to your writing profile, the same way it has access to your calendar or docs.
The profile is readable Markdown, so you can inspect it, edit it, export it, and use it with Claude, GPT, Gemini, Ollama, or inside the Noren Mac app / Chrome extension.
I’d love feedback on three things:
- Does “portable voice profile” make sense immediately?
- Would you trust a tool like this with your writing samples?
- Which workflow feels most urgent: fiction, newsletters, email, LinkedIn, founder updates, or agent-written communication?
Site: https://usenoren.ai
Couple of points - One your website, drive home the point what is the problem and why does "Role" in AI LLM prompts does not work. That is the key piece missing there.
Secondly in your FAQ, there are a lot of technical terms aimed at people who understand what is inference or extraction but this maynot be the language spoken by your users or paying customers. It would be better to talk to them in their language.
Wow that's really funny, I had the same idea related to voice extraction but targeted an other population... I ended up pausing it, but well, maybe I should get back to it haha
Anyway, good luck with Noren xD
Trust me, it hasn't been easy lol. But we keep working on it.
The romance writer vs newsletter writer split is the most interesting part — those two groups want almost opposite things from "voice preservation," even though the surface request looks the same. I've been building a tiny iOS memo app solo (a Captio replacement) and ran into a similar shape: my first ~40 users split into "send-to-email" people and "personal journal" people, and the moment I tried to serve both equally, the thing that made it worth opening got diluted. One thing I'd try: ship the same product behind two different landing pages and onboarding copy, and see which lane converts at a different rate. Are romance vs newsletter folks using it in the same single session, or are they actually bringing in totally different reference samples?
That is worth testing and we started testing it recently. We have landing pages for a couple of use cases but organic traffic hasn't yet ticked up for any.
I think the problem is clear.
Most AI writing tools can make clean text, but they also remove the person behind the writing.
“Portable voice profile” makes sense, but I feel “personal voice layer” or “writing identity” is easier to understand.
For use case, I think founders and people who write online often may feel this pain more than fiction writers.
The simple promise could be:
“Give your AI tools your real writing voice.”
That explains the value very fast.
Agreed! tbh this was built first for founders. But our first sets of users have been fiction writers and students. Maybe we are not doing a good job reaching founders or they are not interested in setting up yet another ai thing.
“Portable voice profile” makes sense, but I think the stronger phrase is probably closer to “writing identity” or “personal voice layer.”
The trust problem is not just whether people will upload samples. It is whether they believe the tool will preserve what makes their writing feel like theirs instead of smoothing it into generic AI prose.
That matters even more for agents. Once an agent writes emails, replies, founder updates, or negotiations on someone’s behalf, the product stops being a writing helper and becomes an identity layer.
Noren is clean, but I’d be careful that it does not make the product feel smaller than that. A name like Lyriso.com would probably fit the human voice/writing identity angle better if you want the brand to feel more ownable and less utility-like.
Thanks for the feedback.
Of course.
If Noren stays as a writing helper, the name is fine.
But if it becomes the personal voice layer agents use to write on someone’s behalf, I’d pressure-test the brand early.
That category needs more trust than a normal writing tool.
Noren means identity and our whole idea is about presserving your identity in the machine. I appreciate you validating our brand direction.
Makes sense.
If the meaning behind Noren is identity, then the direction is aligned.
The main thing now is making sure users understand that meaning fast enough without needing it explained.
If “identity in the machine” becomes clear in the first read, the name can work.