frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Open in hackernews

Show HN: An API for word timestamps from speech, 75% within 20ms

https://quebec.yarn.so
4•jasperstory•1d ago
Hey HN, we're Yarn (W24) and we make a video editing app for startups, and we built Quebec (https://quebec.yarn.so), an API that gives you accurate word timestamps from speech. We needed it because cutting a word even 50ms off from its true timing can ruin talking head footage (examples in the write up).

Accurate word timestamps usually come from a forced aligner, which takes the speech audio and a transcript and works out exactly when each word starts and ends. We weren't satisfied with the accuracy of Montreal Forced Aligner (MFA) or new models like ElevenLabs Cue (https://elevenlabs.io/blog/cue-aligner). Giving LLMs access to a waveform/spectrogram didn't help much either.

So we built our own. There are two particularly interesting parts: - We wrote our own decoder around MFA's acoustic model that only lets each word land within 200ms of the rough timings. That made search small enough to solve exactly, so we don't need to use MFA's beam search. - We added a small CNN, built on the first two layers of Whisper Tiny, that nudges each word boundary by up to 50ms. We couldn't find enough hand-labelled speech that allowed for commercial use, so we had Gemini label our own (17k word boundaries). It took a couple of hours and cost under $200. A net trained on Gemini's labels scored the same as one trained on human labels (73.9% vs. 74%).

One way to measure accuracy is % of word boundaries within 20ms of hand labels. Here are the results on Buckeye (hand-labelled American speech that we didn't train on):

- Quebec 75%

- MFA 61%

- ElevenLabs Cue 30%

If you want to try it, it's free until tomorrow night. You can point your agent at our docs at https://quebec.yarn.so, and send audio on its own or with your own transcript. After that, it costs $0.32 per hour with no signup (minimum $10 top-up). We serve it at cost. Caveats: only tested in English and slower than ElevenLabs.

Let us know if Quebec works well for you, and send us any samples that work badly (you can email support at yarn dot so). Also let us know if you've built or found a better aligner. In theory we would just switch to a better aligner, although we would probably just push a bit harder to make ours better.

Thanks!

Jasper

Gemini 4 Neon

https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-neon/
726•yigitdemirag•10h ago•611 comments

Unauthenticated remote code execution in Nginx

https://www.wiz.io/blog/nginx-http2-rce-vulnerabilities
1301•speedping•16h ago•570 comments

15M web servers on one VM – smoltcp, DPDK, AF_XDP

https://kapteko.com/15-million-webservers/
36•fessguid•1h ago•7 comments

Multiple vulnerabilities in Regexp and String case conversion

https://www.ruby-lang.org/en/news/2026/11/18/multiple-vulnerabilities-in-regexp-and-string/
205•runesoerensen•7h ago•68 comments

Seven new models at scalable price-performance

https://microsoft.ai/news/microsoft-ignite-2026/
517•gmays•15h ago•342 comments

Anthropic prices initial public offering

https://www.anthropic.com/news/anthropic-prices-initial-public-offering
158•surprisetalk•8h ago•160 comments

Axelera Titania Brings Native FP8 and FP4 to In-Memory Compute

https://futurumgroup.com/insights/axelera-titania-brings-native-fp8-and-fp4-to-in-memory-compute/
21•muxamilian•2d ago•0 comments

Inside Deutsche Bank

https://www.inside-db.com/
651•gyre007•15h ago•208 comments

LLM assisted reverse engineering

https://www.slowchronicle.com/posts/llm-assisted-reverse-engineering
166•ioma8•9h ago•44 comments

Genie 4 is now available

https://blog.google/innovation-and-ai/models-and-research/google-deepmind/genie-4-now-available-p...
236•leumon•12h ago•160 comments

DaisyRose-27B

https://www.radixark.com/blog/daisyrose-27b
187•de_sty•11h ago•59 comments

Dumping an AES key from a MyQ garage opener to stop rebooting when offline

https://kusha.me/blog/stop_myq_homebridge_819lmb_rebooting/
107•kusha•2d ago•36 comments

Camera Index

https://www.cameraindex.co.uk/
21•bookofjoe•1d ago•6 comments

Mosquitoes are Ladakh's latest tourists

https://ruralindiaonline.org/article/mosquitoes-are-ladakh-s-latest-tourists
37•akbarnama•1d ago•1 comments

Show HN: An API for word timestamps from speech, 75% within 20ms

https://quebec.yarn.so
4•jasperstory•1d ago•0 comments

The Cambridge Dictionary has named ‘human’ as the Word of the Year 2026

https://www.cam.ac.uk/news/cambridge-dictionary-reveals-word-of-the-year-2026
269•DeepLogin•16h ago•194 comments

Logitech Harmony IR code online database archive

https://github.com/pickysysadmin/logitech-harmony-ir-archive
8•cf100clunk•2d ago•1 comments

Unreal Tournament 2004 in the Browser

https://ut2004.dev/
145•TheAceOfHearts•6h ago•59 comments

Palo Alto Networks to Acquire Applied Compute for $4.5B

https://investors.paloaltonetworks.com/news-releases/news-release-details/palo-alto-networks-acqu...
52•alpb•7h ago•11 comments

Advancing chip design with OpenAI models

https://openai.com/index/advancing-chip-design-with-our-models/
200•vighneshiyer•13h ago•130 comments

Show HN: ExitPath – Open-source ways out of Slack, Notion and Microsoft 365

https://exitpath-ruddy.vercel.app/
202•zerocool0101•15h ago•61 comments

We need smart people who give a shit

https://violetpixel.com/blog/2026-11-17-we-need-smart-people-who-give-a-shit/
469•ExMachina73•7h ago•239 comments

F-22 Next-Generation micro–Data Transfer Cartridge (mDTC)

https://sam.gov/workspace/contract/opp/8948e8601e964c6aba9c801eee16b1be/view
167•birdboat00•16h ago•109 comments

Show HN: Take Voyager 1 apart, down to a single memory chip

https://rshekhar.com/voyager/
184•raj-shekhar•15h ago•46 comments

Show HN: Open-source, real-time object tracking for live video feeds

https://github.com/aryamandarda/tamserve
78•aryaman_darda•2d ago•3 comments

Gemini 4 Neon looks like an Opus 5.5 competitor

https://micjm.com/blog/gemini-4-neon-looks-like-a-top-model/
211•seizethecheese•7h ago•67 comments

Cuttlefish on Mac: Android's Virtual Device Framework on Apple Silicon

https://engineering.fb.com/2026/11/13/android/cuttlefish-on-mac/
63•acossta•3d ago•12 comments

AI agents ask better questions than KPMG accountants in audit trial

https://europeanbusinessmagazine.com/ai-agents-better-questions-kpmg/
442•Betelbuddy•15h ago•192 comments

DNS resolvers are the wrong place to fight piracy

https://techstackups.com/articles/quad9-leaves-france-dns-blocking/
653•ritzaco•18h ago•263 comments

Launch HN: Maritime (YC F26) – A managed cloud for production AI agents

52•ashriv•14h ago•47 comments