[HN Gopher] Launch HN: Cuckoo (YC W25) - Real-time AI translator...
       ___________________________________________________________________
        
       Launch HN: Cuckoo (YC W25) - Real-time AI translator for global
       teams
        
       Hey HN! We're Yong Hee and Gunwoo from Cuckoo (https://cuckoo.so),
       a real-time AI translator for global sales, marketing, and support.
       Companies like Snowflake and PagerDuty use us in Zoom calls and in-
       person meetings, including for technical discussions. Here's a
       demo: https://www.youtube.com/watch?v=LKU1fj3rxh4.  With Cuckoo,
       you get an always-available, real-time translator that adapts to
       your conversations, helping teams avoid wasted time from language
       frictions.  Language barriers are still a big challenge for
       companies with offices in different countries. I experienced this
       firsthand at my previous company in Korea when I invited
       international speakers to our meetups. Much of the nuance and
       technical details were lost, even though many attendees spoke
       adequate English. Speakers often felt isolated, unable to
       understand Korean discussions, and missed out on opportunities to
       network and market their products--despite these events being
       organized for customer interactions.  Meetings with interpreters
       are often twice as long, and details are lost in translation.
       Sometimes a bilingual colleague steps in as an interpreter, but
       they'd be far more valuable focusing on their core roles. If you
       are a sales engineer or customer success engineer, this gets worse
       with highly complex and technical topics.  While LLMs work well for
       translating written text, they struggle with real-time spoken
       translation. For example, OpenAI's Realtime API only handles turn-
       by-turn interpretation, which doesn't work for dynamic
       conversations. One of our customers used to record meetings with
       OBS, upload the recordings to an AI notetaker, and wait for
       translated notes--but that's hardly real-time.  Cuckoo works by
       inviting a bot to your Zoom meeting. You can contextualize
       translations by uploading relevant documents like pitch decks or
       product manuals. You can share the live transcript with your
       customers and teammates with a link. It also works in person.
       We've been surprised by the range of use cases. Snowflake uses
       Cuckoo to host international customers in their Customer Experience
       Centers and Weights & Biases runs VIP roundtables in Seoul. VP of
       Sales at a high-growth startup closed enterprise deals in Japan and
       an early-stage founder in SF connected with their first user in
       Turkey.  Many of our users include engineers who face language
       barriers in technical conversations. For example, a solution
       engineer at PagerDuty uses Cuckoo to troubleshoot integrations with
       local partners in Korea and Japan. These conversations are often
       full of technical jargon, APIs, and organizational context. By
       uploading relevant documents like API docs, they ensure
       translations are accurate and context-aware, reducing
       misunderstandings and inefficiencies.  On Privacy and Security:
       like human interpreters, we sign NDAs to protect sensitive
       information. Transcripts can be set to self-destruct after
       meetings, and we're working toward ISO 27001 and SOC I/II
       certifications. If you delete your account, all associated data is
       erased immediately and irreversibly.  You can sign up with just an
       email and name. After signing up, you can copy and paste your Zoom
       or MS Teams link to have Cuckoo join your meeting. We provide a
       quick 60-minute free trial, but if you need more, please let me
       know at yonghee@cuckoo.so. We'd be happy to extend it for HN
       users!.  The best way to understand Cuckoo is to try it out for one
       of your upcoming global meetings. It takes <3 minutes to get
       started. We're looking forward to hearing your thoughts and
       feedback in the comments!
        
       Author : yonghee
       Score  : 40 points
       Date   : 2025-03-03 18:39 UTC (4 hours ago)
        
       | joshdavham wrote:
       | Congrats on the launch!
       | 
       | Also question: your writing makes you seem quite bilingual and
       | fluent in English. Given this, would you consider yourself a user
       | of your own product? Do you often find yourself needing to use
       | it? It strikes me that the main users would be people who
       | struggle with English specifically. Though I guess with recent
       | innovations in China, potentially more English speakers will
       | start needing to translate from Chinese.
        
         | yonghee wrote:
         | I love this question!
         | 
         | Yes, I am bilingual. I was fortunate enough to study both in
         | Korea and Canada.
         | 
         | I use our product every day when I'm meeting with customers in
         | Japan and China. We joke that we are our very first customers.
         | Personally, it's best when I get to meet them in person and use
         | our in-person meeting feature since I get to see their
         | reactions.
         | 
         | I would say half of our users are fluent in English since they
         | mostly work for U.S. companies. The other half would be people
         | in Korea, Japan, China, and more who need the language support.
        
       | dleeftink wrote:
       | Love it! Friction is part of (language) learning, so hopefully
       | some doses will remain down the line.
        
         | yonghee wrote:
         | Cuckoo x Duolingo - the best combo!
        
       | donbox wrote:
       | Congrats on the launch! Does it support Hinglish (Hindi +
       | English)? I'm a native Hindi speaker, but I found the Hindi in
       | the demo video hard to follow. Many Hindi speakers, especially in
       | everyday conversations, naturally mix Hindi and English -- often
       | using English for technical or difficult terms. Would love to
       | know if the model handles that kind of language blend
        
         | joeevans1000 wrote:
         | Plot twist!
        
         | yonghee wrote:
         | My batchmate was saying the same thing, actually. He was trying
         | to use only Hindi for the demo sakes and it was almost
         | difficult for him to explain his product without referring to
         | English.
         | 
         | We do our best to deal with langauge changes. For example, when
         | talking bio, almost half of the sentence is in English terms
         | and Cuckoo does pretty well in that context as well!
        
       | aresant wrote:
       | This is awesome!
       | 
       | Is there a consumer version available?
       | 
       | Or is there a company focused on that side of the business?
        
         | yonghee wrote:
         | While we are focusing on business use cases, we are seeing a
         | few invidivuals sign up for their own uses.
         | 
         | Email me at yonghee@cuckoo.so so I can help you out with first
         | few months!
        
       | mandeepj wrote:
       | I see you are using these two words - Interpreter and Translator
       | - interchangeably! They aren't same; there's a big difference
       | between them [0].
       | 
       | From your demo, I gather you are a translator, which is a big
       | let-off for me. Reading and understanding text is much slower
       | than just listening. Also, spoken words are just 30ish% of the
       | overall communication. I'm afraid while your users would be busy
       | in reading translated text, they'd lose out on other vital
       | communication cues like hand gestures, facial expressions etc.
       | 
       | Is real-time audio interpretation in the pipeline?
       | 
       | [0] -
       | https://www.google.com/search?q=translator+vs+interpreter&oq...
        
         | givemeethekeys wrote:
         | > Reading and understanding text is much slower than just
         | listening.
         | 
         | While watching their demo video, I had no trouble reading and
         | interpreting the translated English at the speed the
         | conversations were going on. There's a chance that some
         | speakers would speak much quicker, but I think this software
         | covers the vast majority of use-cases.
         | 
         | Real-time translation is a great start. I'm sure these models
         | can be tweaked over time for better interpretation, especially
         | given that they learn based on context.
         | 
         | There is another aspect to this: the small pauses forced by
         | technology will give people just enough time to think, which is
         | welcome in a business meeting.
         | 
         | Full disclosure: I am not a user or a customer, but this looks
         | like it is something I would one day want to use if the
         | opportunity presents itself.
        
           | yonghee wrote:
           | I think it's similar to Netflix subtitles--some people prefer
           | subtitles and dislike voice-overs, while others opt for the
           | dubbed versions.
           | 
           | I also believe that as the meeting progresses, it feels more
           | natural, and participants become aware of the translator.
           | (Interestingly, they often start speaking more clearly and
           | using fuller sentences, just as they would with human
           | interpreters!)
           | 
           | Thanks for your comment. I hope you give it a try!
        
         | carlosjobim wrote:
         | In a business setting, the only communication that has any
         | legal standing is the spoken word and the written word.
        
           | yonghee wrote:
           | That's a great point. An added benefit of our approach is
           | that it provides a written version of the conversation in
           | multiple languages.
           | 
           | (Of course, some users may prefer to remove the conversation
           | entirely for data security and privacy reasons.)
        
           | mandeepj wrote:
           | > In a business setting, the only communication that has any
           | legal standing is the spoken word and the written word.
           | 
           | Legal, huh? How many indictments have you seen come out of a
           | business meeting? The expression representation is very much
           | part of a work environment. Someone may not say a word while
           | hearing a crazy idea, but they'll certainly roll an eye.
        
         | yonghee wrote:
         | I agree with your observation about "Interpreter" vs.
         | "Translator."
         | 
         | When we first started this project, we referred to it as an
         | "interpreter." However, after speaking with human interpreters
         | and considering their feedback, we settled on "real-time
         | translation". We might have left some of our past on the
         | internet tho..
         | 
         | As with everything, there are both advantages and limitations
         | to text-based translations. Here are a few:
         | 
         | Limitations:
         | 
         | - Some people may find it challenging to follow gestures and
         | expressions while reading.
         | 
         | - In more one-way scenarios, such as presentations and
         | webinars, hearing the speaker's voice often feels more natural.
         | 
         | Pros:
         | 
         | - Many users actually prefer text because it allows them to
         | hear the speaker's original voice and pick up on nuances.
         | 
         | - Having a written record enables post-meeting summaries and
         | the opportunity to repurpose transcripts into other materials,
         | such as blog posts, custom user manuals, JIRA notes, and more
         | using AI.
         | 
         | - There are also technical constraints with voice-to-voice
         | translations, which currently tend to be turn-based rather than
         | real-time (streaming) - not ideal for exchange of ideas.
         | 
         | That said, we are excited to see how the TTS and STT
         | technologies evolve and are looking forward to experimenting
         | with "interpretation" in the future!
        
         | bityard wrote:
         | > Reading and understanding text is much slower than just
         | listening.
         | 
         | Speak for yourself! I read _much_ faster than listening to
         | someone saying the same thing. This is why I can't stand
         | subtitles on videos, movies, and tv shows. Because of how my
         | brain works, I can't help but read the text. And when it's
         | there, I'm done reading the person's line when they are only
         | 25-50% through speaking it. So it "feels" like I'm watching a
         | show where everyone repeats the last half of every sentence.
         | 
         | > Is real-time audio interpretation in the pipeline?
         | 
         | When I saw the headline, I assumed the product was doing real-
         | time translation and voice cloning in one. Now _that_ would be
         | an interesting use of AI. (Google and others have been doing
         | real-time voice recognition and text translation for years.)
        
       ___________________________________________________________________
       (page generated 2025-03-03 23:00 UTC)