News for hackers. Stuff that matters.

louisjoejordan

Show HN: Stem splitting and audio conversion running in-browser with Rust/WASM

https://www.audiodrop.app/

I got tired of all the ugly "free" audio conversion tools littered with ads so I built my own that runs entirely on your machine in the browser.It does stem separation (vocals/drums/bass/other), format conversion, BPM detection and speed/pitch changes. Nothing is uploaded (there is no api) + the models runs client-side.Stack is Rust compiled to WASM, mediabunny for conversion +...

( Read More... | no comments | on HN )
Re:ElevenLabs Music v2.5 on Friday September 11, 2026 @07:44PM (#49666965)
1. Hopefully this should be fixed for you now 2. We're working very hard to get those 2 numbers aligned
Re:ElevenLabs Music v2.5 on Friday September 11, 2026 @07:34PM (#49666888)
I came up with the name Muse before Meta did so this was painful
Re:ElevenLabs Music v2.5 on Friday September 11, 2026 @07:30PM (#49666851)
Which part do you hate the most?
Re:ElevenLabs Music v2.5 on Friday September 11, 2026 @07:27PM (#49666825)
I'll take perfectly listenable
Re:ElevenLabs Music v2.5 on Friday September 11, 2026 @07:27PM (#49666814)
Thank you!
Re:ElevenLabs Music v2.5 on Friday September 11, 2026 @07:26PM (#49666812)
Curious to know what you'd like to see in this space ^
Re:ElevenLabs Music v2.5 on Friday September 11, 2026 @07:24PM (#49666794)
Engineer on ElevenMusic (so discount accordingly). I like to think of the model as an instrument - something you can sit with for hours and get better at. The people who stick around the most treat it that way (they write their own lyrics, upload half finished songs as reference, bring their own voice back from old recordings etc..). For example my dad was in a band in the 80s and stopped writing music and it's fun to see him make it again.

Every time a barrier drops (samplers, synths, autotune, GarageBand, sample packs) this argument gets made and it is partly right. Plenty of what gets made on our platform is complete slop but if your only contribution to a song was a "make a pop song" prompt that's what you deserve. You are right that nobody was ever short of music - but many people are short of a way to make it so we build for them.

Re:ElevenLabs Music v2.5 on Friday September 11, 2026 @06:45PM (#49666441)
Same here - the pieces exist & our music model can now do voice + sfx + music in a single generation (5 min limit so these were chained together to form longer pieces).

I made some here with our last model: https://elevenmusic.io/artists/elevenmusic-spaces

Re:ElevenLabs Music v2.5 on Friday September 11, 2026 @06:12PM (#49666117)
I'm an engineer on elevenmusic and we are working hard to restore availability (which has been sporadic over the last 10 minutes). We're working very hard behind the scenes to bring it back online.
Very nice!

Show HN: I tried to build the first AI radio station

https://www.cube.fm/

Hey HN,I grew up in Luxembourg & spent years near the old radio tower that blasted pirate radio across borders (unlicensed DIY broadcasts). I’ve always been into radio & wanted to hear what a 24/7 AI-generated “station” might sound like. It’s powered by ElevenLabs. GPT-5 gives each artist a little personality & generates music in their style.I built a simple algorithm that tracks...

( Read More... | no comments | on HN )
Re:Show HN: ElevenLabs UI shadcn/UI components for audio on Tuesday October 07, 2025 @06:36AM (#45501442)
Try hitting the sparkley button at the top!

Show HN: 11.ai – Talk to Hacker News with your voice (reads comments)

https://11.ai

Hi HN,I built 11.ai, a proof-of-concept voice assistant that can read Hacker News threads aloud.It’s powered by the Model Context Protocol (MCP) plus ElevenLabs Conversational AI, so it can also take practical actions inside the tools you already use.It's free to use, excited to hear your feedback.

( Read More... | no comments | 3 points | on HN )
Re:Eleven v3 on Thursday June 05, 2025 @04:30PM (#44195490)
quick note that that voice selection matters a lot with our new v3 model, especially voice language!

We have a curated list of v3 voices in the library, but feel free to try others to find what works. Make sure language <> voice language match.

Re:Eleven v3 on Thursday June 05, 2025 @04:27PM (#44195468)
Might take a few tries, but it will.
Re:A guide on how to prompt engineer voice AI agents on Saturday March 29, 2025 @10:56AM (#43516043)
I put together a guide on prompt engineering for voice agents. It breaks down six core elements. Feedback welcome, what have you found works best?
Apologies sir
We should be releasing the voice next week.
Very strange - let me look into this. Thanks for flagging.
Thanks for the feedback - What do you mean by the censor? As in safety moderation? For this specific agent I had to prompt it very carefully to makes sure people wouldn't ask Santa for anything dangerous :)

Billing component is noted, round-up doesn't seem fair. Let me check this.

Thanks! We created a custom professional voice clone for this purpose using a voice actor. Let me check with Santa to see if we have permission to make his voice public :)

Older »