This product was not featured by Product Hunt yet. It will not be visible on their landing page and won't be ranked (cannot win product of the day regardless of upvotes).
Lokutor
Voice AI on speech foundation models we train ourselves
Lokutor is a speech AI company in Madrid. We design and train our own small speech foundation models: Versa (text to speech in 9 languages, incl. Catalan, Galician and Basque), Turno (turn-taking: knows when the caller is done and lets them interrupt) and Psst (noise suppression), with Conv 2.0 (speech recognition) in development. Small models make it fast and affordable: agents answer in about 0.6 s (596 ms median) for about 5¢ a minute, LLM included. Live demo on lokutor.com, no sign-up.
Hi Product Hunt, I'm Daniel, founder of Lokutor in Madrid, with my co-founder Guillermo.
We got into voice AI because every voice agent we tried was a stack of other people's models, with their latency and their prices. So we design and train our own: Versa for the voice, Turno for knowing when you've finished speaking, Psst for noise, and Conv for recognition (today we run NVIDIA's open Parakeet on our own servers while Conv 2.0 gets ready). They are small, which is why the same stack can run in our EU cloud, on a customer's own servers, or on a device.
What you get today: voice agents that start answering about 0.6 s after you stop talking (596 ms median over 30 production calls on 10 October, details at lokutor.com/benchmarks), phone numbers, tools and CRM, text to speech in nine languages including Catalan, Galician and Basque, speech to text and noise cancellation as APIs, and voice cloning. A typical agent minute is about 5¢ with the language model included, billed by the second from a prepaid balance. No plans.
Try the voice demo at lokutor.com, no sign-up. New accounts get $5 of credit with no card, and for launch week there is $25 more with the code PRODUCTHUNT (30 days to use it, first 500 accounts, until 20 October): app.lokutor.com/redeem/PRODUCTHUNT
If you build hardware: Ito and Oído are our smallest models, open source, for the ESP32 S3 (physical-board performance is still being measured). If you need on-premise or more concurrent calls, lokutor.com/enterprise comes straight to us.
I'd love to hear what breaks, what's missing, and which language or voice you need next.
Hey, great work; just doing some light testing but seems like a solid product so far. If anything I would devote some extra time to fine-tune the expressivity of the voices, which (and maybe that's just me) even with the settings at max, they feel a little too flat. Also, and I'm aware this is something everybody struggles with, see what can you do with commas, semicolons, etc. as the pauses and intonation after them are still a little off. Please take these comments as what they are, just suggestions to ensure you deliver the very best, not a hard criticism at all! Keep up the good work!
Hey, congrats on the Lokutor launch! We're sharing a launch day with Benson: https://www.producthunt.com/products/benson-2. Would appreciate a comment or a little support if you have time today. Wishing you a great launch!
congrats on the launch!
BTW I was trying the catalan voice chat by choosing catalan in the dropdown but it kept talking in spanish, i don't know if there's any setting i'm configuring wrong!
About Lokutor on Product Hunt
“Voice AI on speech foundation models we train ourselves”
Lokutor was submitted on Product Hunt and earned 21 upvotes and 9 comments, placing #31 on the daily leaderboard. Lokutor is a speech AI company in Madrid. We design and train our own small speech foundation models: Versa (text to speech in 9 languages, incl. Catalan, Galician and Basque), Turno (turn-taking: knows when the caller is done and lets them interrupt) and Psst (noise suppression), with Conv 2.0 (speech recognition) in development. Small models make it fast and affordable: agents answer in about 0.6 s (596 ms median) for about 5¢ a minute, LLM included. Live demo on lokutor.com, no sign-up.
Lokutor was featured in API (98.8k followers), Developer Tools (521.6k followers) and Artificial Intelligence (481.7k followers) on Product Hunt. Together, these topics include over 233.1k products, making this a competitive space to launch in.
Who hunted Lokutor?
Lokutor was hunted by Daniel Varela. A “hunter” on Product Hunt is the community member who submits a product to the platform — uploading the images, the link, and tagging the makers behind it. Hunters typically write the first comment explaining why a product is worth attention, and their followers are notified the moment they post. Around 79% of featured launches on Product Hunt are self-hunted by their makers, but a well-known hunter still acts as a signal of quality to the rest of the community. See the full all-time top hunters leaderboard to discover who is shaping the Product Hunt ecosystem.
Want to see how Lokutor stacked up against nearby launches in real time? Check out the live launch dashboard for upvote speed charts, proximity comparisons, and more analytics.
Daniel Varela
Hi Product Hunt, I'm Daniel, founder of Lokutor in Madrid, with my co-founder Guillermo.
We got into voice AI because every voice agent we tried was a stack of other people's models, with their latency and their prices. So we design and train our own: Versa for the voice, Turno for knowing when you've finished speaking, Psst for noise, and Conv for recognition (today we run NVIDIA's open Parakeet on our own servers while Conv 2.0 gets ready). They are small, which is why the same stack can run in our EU cloud, on a customer's own servers, or on a device.
What you get today: voice agents that start answering about 0.6 s after you stop talking (596 ms median over 30 production calls on 10 October, details at lokutor.com/benchmarks), phone numbers, tools and CRM, text to speech in nine languages including Catalan, Galician and Basque, speech to text and noise cancellation as APIs, and voice cloning. A typical agent minute is about 5¢ with the language model included, billed by the second from a prepaid balance. No plans.
Try the voice demo at lokutor.com, no sign-up. New accounts get $5 of credit with no card, and for launch week there is $25 more with the code PRODUCTHUNT (30 days to use it, first 500 accounts, until 20 October): app.lokutor.com/redeem/PRODUCTHUNT
If you build hardware: Ito and Oído are our smallest models, open source, for the ESP32 S3 (physical-board performance is still being measured). If you need on-premise or more concurrent calls, lokutor.com/enterprise comes straight to us.
I'd love to hear what breaks, what's missing, and which language or voice you need next.