We already are getting a overwhelming response in ourr website and product and people are loving it, try it out first hand
https://platform.vaani.dev/
Grab your free credits, and this thread and our socials are open for feedback and comments, Thank you so much!

Our callers mix Hindi and English in the same breath: "haan toh mein bol raha tha, why are you calling me?" We assumed language handling was a detection problem. It turned out to be about six separate problems, and we found them one at a time, in production.

1. Detecting a language switch without flip-flopping
A caller says one English word and the bot jumps languages. We settled on a heuristic switch with guardrails you can configure per agent:

• a minimum number of words per turn (default 3)

• the same language for consecutive turns (default 2)

• a dominant-language threshold (0.6) from a fastText classifier

• a reset of the counter on mismatch, so one stray turn doesn't count

• "meta" references ignored, for example someone saying "English" while speaking Hindi

2. Callers who just tell you what they want
"Let's speak in English" shouldn't need two confirming turns. We added explicit request detection, in English and Hindi (e.g. " "). It bypasses the heuristics and switches immediately.

Here's how webhooks work in Vaani (no code needed)

A webhook is just Vaani sending your app a message when something happens on a call.

You pick the events you care about: call_started, call_ringing, user_picked_up_at, call_rejected, call_no_answer, call_failed, call_ended.

You give us a URL. When a call ends, we send it a message, and you can drop the result in a Google Sheet, ping Slack or update your CRM.

Really excited to take our best product yet with you and have best product people experience it first hand.
Avail 1000 platform credits, exclusive just for PH users by using the code PHL1000X.
website: https://www.vaani.dev/
Platform link: https://platform.vaani.dev/
The comments and this thread is open for everything about the product, thank you so much for trying out the platform!

as an engineer at Vaani, i've spent a good chunk of time building the stuff that makes voice agents actually work in production.

now we've opened our platform to everyone.

build voice agents, design call flows, integrate tools, connect knowledge bases and go live on a real phone number, without writing code.

devs, PMs, builders: what would you build if voice AI wasn't a pain to set up?

One of the most frequently asked questions I get is whether you would implement s2s because of its lower latency and costs.

While it offers lower costs and lower latency/quicker responses, it doesn't provide the same level of configurability and control that we all expect. While foundational models make the voice agent sound like a human, a cascaded architecture allows it to behave like one too.

While building Vaani, one thing we spent quite some time getting right was call transfers.

It s not always about transferring a call to a human. Sometimes the conversation needs to move to a different, more specialised agent. Other times, it needs to go to a human who can take over.

Making these transitions feel natural is harder than it sounds. You don t want the conversation to lose context, the caller to repeat themselves, or the whole experience to feel like a disconnected call.

We ve worked on making these handoffs smooth at Vaani, including agent-to-agent and agent-to-human transfers.

Ever had to explain your missing lunch to a support bot... for the 4th time?

You call. You explain. You hang up. You call again, and the bot greets you like a total stranger. Hi! How can I help you today? We ve been here. (On an empty stomach, no less).

We hated this too. So at Vaani, we built Memories: our voice agents remember your past calls and what went wrong, so every conversation picks up right where the last one left off. No repeat sob stories. No pulling your hair out. Just agents that actually get it.

Building voice agents? Give your users the gift of never having to say as I told you last time... ever again.

Hey Product Hunt!
We re live with the Vaani developer platform, the only voice AI harness that you ll ever need.

Building a demo agent takes a weekend. Keeping it alive on real calls is the hard part. A few things we keep hearing in the Voice AI community -

• the model hallucinates or fails mid-call (we built automatic failover to a backup model for this, so the call doesn t die)

• callers go silent, interrupt, or throw objections nobody planned for ( we have interruption handling, filler words, and background noise to tackle that)

• nobody wants to read 500 transcripts to find out what happened ( we have summaries, extractions, and

so we re curious about your experience:

Scroll through Product Hunt for two weeks, and you'll see a new voice AI agent, dubbing tool, or "AI receptionist" almost daily. As somebody who's been building in this space and observing it for the last 2.5 years, here're a couple of insights I gathered over time:

Why so many?
Voice is the one interface nobody has to learn. We spoke long before we wrote. Keyboards, apps and dashboards all needed training, but talking doesn t. So once quality got good enough post-2022, adoption didn t need a change in behaviour. It only needed a threshold to be crossed across latency, naturalness or say, interruption handling.
The demand was also already sitting there. A lot of business still runs on phone calls: support, collections, appointments, follow-ups. They re expensive to staff and hard to scale. Also, during last 2 years, building demos got cheaper and cheaper. With decent APIs for speech, LLMs and telephony, one person can wire up a convincing agent in a weekend. This is good for innovation and building use cases for something as niche as HVAC service renewals support.
But what's good for innovation is also the reason for the noise. Buyers are the ones who pay for that noise. Every two-minute demo looks the same. Everyone says 'human-like' and 'enterprise-ready', and almost none of it can be checked before you pay for a pilot. The cutthroat market is trying to undercut each other's prices to zero.