Real-time Voice AI. Built for Scale.

San Francisco
A huge thanks to our customers, partners, investors, and the team. We're just getting started.
We have passed the Turing test for Voice AI and raised $13M in Series A funding to bring the next generation of Voice to enterprises.
3
1
32
9,899
smallest.ai retweeted
POV: CTO and CEO after closing a 7 figure deal
21
15
200
19,264
The real test of voice AI is what happens in production. Yesterday at @NICELtd World Mumbai, people ran that test themselves at our booth. As Knowledge Partner at this year's @NICELtd World, we had @apoorvsood, our Head of GTM, on stage to walk through what real-time voice AI looks like inside an enterprise. Next stop, Delhi. Tomorrow, 25 September, at Roseate House, Aerocity, from 3:30 PM.
3
3
14
761
The real test of voice AI is in production. Platforms have gotten very good at deciding what to do next. Speaking that decision aloud to a customer, in real time, is still where most stacks fall apart. That layer is what we build. On 23 September we're at @NICELtd World Mumbai as Knowledge Partner. @apoorvsood, our Head of GTM, takes the stage on real-time voice AI for enterprises. If you run CX or contact centers and you're deciding what your voice stack should look like, come say hi.
2
1
12
489
At @GffFintechfest, we handed out money. Everyone at GFF thinks about money all day. We designed the brochure as a currency note and called it The Smallest Issue. Inside, it runs like a newspaper, problem, proof, product. Most people didn't clock it was a pitch until somewhere around the benchmarks. The only brochures that survive are the ones people wanted. We build our products the same way.
1
2
9
504
smallest.ai retweeted
A video editor solved his own problem. Now 20,000+ people use it. @tanishqgopani makes the videos you see on our socials. Captions were one of the most tedious parts of his workflow: writing them, timing them, and placing them correctly. He got tired of it and built a fix for himself. Then he put it out for anyone to use. It’s called CapKit. It generates animated captions inside After Effects and Premiere Pro. The testimonials people are sharing are better than anything a marketing team could write. The transcription under the hood runs on Pulse, our speech-to-text model. Its transcription is reliable across different kinds of real-world audio, so most captions come through in one pass, with little or nothing left to edit. This is something I keep relearning from the people here: you cannot assign curiosity. You hire people who have it, and then the main job is to not get in the way. If you work with video, try CapKit.
8
3
43
2,502
We guessed people’s ATM PINs at @GffFintechfest. That’s what it took to make them stop. Every year, the hall gets louder: more booths, more banners, more people walking past all of it. One more pitch was never going to cut through. So we gave people something they couldn’t explain. We hired a mentalist. He guessed phone numbers, favorite movies, childhood nicknames, and numbers people had locked in their heads seconds earlier. At one point, he guessed someone’s favorite Kannada movie. He doesn’t speak Kannada. The people around him stopped walking and started reacting. That was the point. At a crowded event, anyone can say more. Making someone stop is the harder job. Then you earn the conversation.
1
9
506
Smallest AI is now integrated with @bolna_dev. If you're building phone-based voice agents on @bolna_dev, you can now run them on Smallest's fastest models: Pulse for STT, Lightning for TTS for real-time, natural conversations over any phone call. Live today! Setup guide below.
1
4
34
2,757
smallest.ai retweeted
Replying to @smallest_AI
@smallest_AI It is top of charts as always!
Inworld Realtime TTS-2 debuts at #2 of 17 on the Voice Arena US English 🇺🇸 TTS Leaderboard at 1068 Elo, above Google DeepMind's Gemini 3.1 Flash TTS and inside the statistical band of Cartesia's Sonic-3.6 at #1. Realtime TTS-2 is Inworld's latest real-time TTS model, built for live consumer applications. The version on the board is the research preview, which Inworld has since taken to general availability with notable improvements on quality, stability, speed and consistency. Every ranking on Voice Arena comes from blind, head-to-head votes by vetted native speakers, scored with Bradley-Terry Elo. Realtime TTS-2 enters a 17-model field at #2 placing it above every model on the board from ElevenLabs, Microsoft, OpenAI and xAI, with clear statistical separation from all four. Results backed by 600+ blind, head-to-head listener votes from native US English speakers. Congratulations to @inworld on the release! See below for the exact audio raters judged and unedited listener verdicts 🧵
1
5
243
smallest.ai retweeted
Mr. @bomanirani ran a photography studio until he was thirty-five. Then he stepped onto a stage. He told us this over dinner in Mumbai last week, during @GffFintechfest week. Twenty-five people from banking and financial services, no stage, no deck - just him, talking about the years before anyone knew who he was. He grew up with a stammer. After a school performance, his mother recorded the applause instead of the singing, and played it back to him whenever he doubted the way he spoke. His first theatre role was small enough that he nearly turned it down. There's a film in his career that was never released. Photographer at thirty-four, stage at thirty-five, commercial cinema at forty-four, director at sixty-five. He is sixty-six. None of it, he said, was time lost - the years that led nowhere were load-bearing. Near the end, a guest asked if he'd rewrite any chapter of his life. Not a single one, he said. Most of us are living through a chapter that looks like a detour right now. You cannot tell which years are the foundation while you're standing on them - you only find that out by not quitting early.
4
34
2,138
smallest.ai retweeted
Exhibition showcases by @smallest_AI , Rezolv AI, @KnightFinTech, Fyno, and @RaiseTheBarHQ presented voice AI, co-lending middleware, and trading architecture to visiting global banking delegations and all attendees.
1
2
4
148
Cashfree is recovering failed payments with @smallest_AI voice agents. When a payment fails, Cashfree Resolve calls the customer within 60 seconds. It explains what went wrong in the language they answer in, then helps them complete the payment. @gocashfree launched Cashfree Resolve at Global Fintech Fest week. On the floor, our Head of GTM, @apoorvsood, was asked: What is so hard about an AI agent explaining a payment failure? Watch him break down the problem below.
3
4
17
1,766
Will voice agents finally replace humans by 2030? This is a question we kept getting asked at Global Fintech Fest, yet again. We’re realising that the answer to this question isn’t just benchmarks or how “real” an agent sounds. It’s trust. For voice agents to replace humans entirely, they need to handle conversations more organically. The moment someone realises they’re speaking to a machine, you lose that sense of connection, and the conversation can end there. Voice agents need to handle the things that make real conversations messy: different accents, different ways of speaking, interruptions, changing context, language switching, and the ability to respond instantly without losing track of the conversation. When agents listen, respond in real time, understand interruptions, and speak the way a person actually would, it changes the conversation. That’s the direction we’re building towards at Smallest AI: voice agents that don’t just complete a call, but make the conversation feel natural enough for people to stay in it. We’re hoping that the next question we get asked isn’t, “Would I let AI handle this?” but, “Why would a human need to handle this?” @GffFintechfest #gff2026
3
3
16
714
Three days of GFF @GffFintechfest, and we barely had a quiet minute. Our booth stayed busy with people trying the voice agents, challenging each other at the game, and getting into conversations with the team. We came to show what we’ve been building and left with a lot more conversations to continue. Until the next one. @apoorvsood @harshja1n @_karandubey @VaibhavMehta277 @GuptaPratirath @tanishqgopani #gff2026
2
19
963