Cerebrium voice agent (fastvoiceagent)
AI
voice agent
SHOW HN2024
Legend
Cerebrium voice agent (fastvoiceagent)
Voice bots with 500ms response times
fastvoiceagent.cerebrium.ai#voice bot#low latency#speech to speech#llm voice#real-time voice ai
315points
99comments
#186of 3,515 in 2024
Top 6%of 2024 by points
How HN took it
MixedNo clear lean
25 positive, 26 neutral or asking, 1 mixed, 9 critical, out of 61 comments about the project. Thread score +0.27.
Performance: 12 for, 3 againstExecution: 4 for, 3 againstDesign: 1 for, 3 againstIdea: 3 for, 0 againstDocs: 1 for, 0 against
The thread, comment by comment
Each row is a top-level comment and its replies, in HN's order. Select a cell to open that comment on HN.
positive neutral or mixed critical about another commenter or something else the maker
What HN said
What people liked
Very fast response, feels conversational
From 4 comments
This thing is incredible. It finished a sentence I was saying.
jaybrendansmith on HNJesus fuck that's fast, and I had no idea speed mattered that much. Incredible. Feels like an entirely different experience than the 5+ seconds latency with openai.
isoprophlex on HNVery, very impressive! It's incredibly fast, maybe too fast, but I think that's the point. What's most impressive though is how the VAD and interruptions are tuned. That was, by…
mdbackman on HNThis was scary fast. Neat interface and (almost) indistinguishable from a human over the phone / internet. Kudos @cerebrium.ai.
realyashnag on HN
Interruption handling is natural
From 2 comments
Very, very impressive! It's incredibly fast, maybe too fast, but I think that's the point. What's most impressive though is how the VAD and interruptions are tuned. That was, by…
mdbackman on HNThis is indeed fast! Also seems to be no issue interrupting it while speaking. Is this using WebRTC echo cancellation to avoid microphone and speaker audio mix ups?
hackerbob on HN
Objections
Voice sounds robotic
From 2 comments
Fast yes, but the voice sounds robotic.
aussieguy1234 on HNIt's literally a robot
cloudking on HN
Interrupts users who pause
From 1 comment
A chatbot that interrupts me even faster. Sorry for the sarcasm. maybe im just slow, but when I'm trying to formulate a question on the spot, I pause a lot. having the chatbot…
SubiculumCode on HN
Firefox support is flaky
From 2 comments
I use Firefox... still.
geofffox on HNHi, I built the client UI for this and... yea, I really wanted to get Firefox working :( We needed a way to measure voice-to-voice latency from the end-user's perspective, and…
makeitmore on HN
Reported latency numbers do not add up
From 1 comment
Your marketing says 500 but your math says 759.
az226 on HN
Requests
Firefox support
From 2 comments
I hate that everyone just develops for chromium only
panja on HNHi, I built the client UI for this and... yea, I really wanted to get Firefox working :( We needed a way to measure voice-to-voice latency from the end-user's perspective, and…
makeitmore on HN
Open-source or self-hostable version
From 1 comment
I too am excited about voice inferencing. I wrote my own Websocket Faster whisper implementation before OpenAI's gpt4o release . They steamrolled my interview coach concept…
trueforma on HN
Questions
Total cost of running the system
From 1 comment
I've been developing with Deepgram for a while, and this is one of the coolest demos I've seen with it! I am curious about total cost to run this thing, though. I assume that on…
c0brac0bra on HN
Which LLM and training data were used
From 2 comments
i have tried it. it is really fast! I know making a real-time voice bot is not easy with this low latency. which LLM did you use? how large LLM to make the conversation efficient?
spark_chicken on HNThat's awesome - can you say anything about what datasets this was trained on? I assume something specifically conversational?
gsjbjt on HN
Alternatives named
Moshi by Kyutai
From 1 comment
moshi by Kyutai seems to have beaten your approach by about 500ms, and they're going to release open source. https://www.youtube.com/live/hm2IJSKcYvo hn discussion here:…
etherealG on HN
GPT-4o voice mode
From 2 comments
It's not exactly clear is this a voice-to-voice model or a voice-to-text-to-voice model? When it is finally released, OpenAI claim their GPT4o audio model will be a lot faster at…
spuz on HNJesus fuck that's fast, and I had no idea speed mattered that much. Incredible. Feels like an entirely different experience than the 5+ seconds latency with openai.
isoprophlex on HN
Siri
From 1 comment
Damned impressive. Apple's Siri still can't allow me to have a conversation in which we aren't tripping over each other and pausing and flunking and the whole thing degrades into…
andrewstuart on HN