... Quick build-in-public update from Sponkly (real-time voice translation for chat, calls and meetings in 32 languages). This week, as a solo bootstrapped founder: NVIDIA Inception: accepted AssemblyAI Startup Program: accepted AWS Activate: $10,000 in credits LiveKit Startup Program: approved today, $17,000 in credits + a year of the Scale plan What I will use it for: Sponkly's calls and meetings are peer to peer today (WebRTC mesh). Great for 1:1, heavy beyond ... ... people. Moving meetings to LiveKit

... side by side or one language), Word, text, subtitles, audio or a link. Who it is for: international students in lectures, conference and trade-fair visitors, teams meeting foreign partners, journalists, language learners. Meetings also moved to LiveKit media this week, so video and sound stay steady with many people. Try it: https://sponkly.com/features/record What would you record first? Comment from Bala Balasubramaniam(@mrbala_subramaniam): the speaker identification feature sounds useful especially for longer meetings. Does it still separate ...

... this year the big one, co-hosting. Friends tap a link and show up on camera in your stream from their own iPhone, with everyone's voice mixed in. Getting that to work meant a lot of ReplayKit, LiveKit and RTMP plumbing, and I'm happy to talk shop about any of it. EveryCord launched here today. If you're building for creators or streamers, I'd like to hear what you're working on and what you've learned ...

... email. The user completes the structured input while they're still talking to the agent, and the response comes back into the conversation. It works as a tool inside your existing voice stack including Vapi, Retell, ElevenLabs and LiveKit. The idea is simple: Voice-first, not voice-only. We're launching @VocaLoop on Product Hunt and would genuinely love feedback from other people building voice agents. What information have you found hardest to collect reliably over voice ...

... languages Interruption Prediction v1 first audio-only model that tells "mhm, keep going" from "wait, stop." Under a second, under 6% false positives Signal Detectors real-time TTS detection, gender, accent classification Already running inside Daily, Vapi, LiveKit, and some of the largest AI labs. 3.5x better turn-taking, 50% fewer dropped calls. All new models bundled at existing pricing. Would love your thoughts. Happy to go deep on any of the technical details in the comments. Comment from ... ... quiet, then responds. That's why every voice agent feels slow and robotic. Our model works differently. It listens to the rhythm and intonation of speech and predicts the end of turn in milliseconds. We tested it against SmartTurn, LiveKit

... hooks But: Those improvements may also pull voice input deeper into apps , creating opportunities for third-party tools to power native voice experiences. Voice Agents are becoming Infrastructure There s a growing ecosystem underneath all of this: @Vapi @Pipecat @LiveKit @Coval This is the voice agent infra layer, a category we ll cover in later Orbit Awards. Comment from Swebliss(@swebliss1): Excellent article. Haven't used an app like it before, I'll try them out! Comment from Aaron ...