Pratik Mundra

SigmaMind AI - Zapier for Voice AI Agents | YC S22

[Pratik Mundra @pratmun]

@GaddipatiHarsha One of the products that I genuinely love! I’m now spending less time on emails. Go team @slashyemail

[Harsha Gaddipati @GaddipatiHarsha]

Excited to announce Slashy

The first email client that works for you.

The real cost of email isn't the time. It's the mental load of constantly checking it, just in case something needs you.

Slashy kills that. You never need to open your inbox unless Slashy tells you.

Try it out at https://t.co/6tuQxWV7KR

[Pratik Mundra @pratmun]

@NamanyayG Congrats on the progress!

[Pratik Mundra @pratmun]

@raidingAI As an early user I can say - best thing I like is @slashyai naturally understands my style and a draft is ready to send as a soon as I open an unread email. Sending emails is so much faster.

[Avinash Raghava 🇮🇳 @avinashraghava]

#53: #AIRadarDaily — @sigmamind_ai

There is a massive illusion in the AI space right now when it comes to voice. The demos look like magic. But the moment you try to deploy a voice agent in a real-world enterprise setting, reality hits. You end up duct-taping together different transcription APIs, LLMs, and text-to-speech models. The result? Unbearable latency, awkward robotic pauses, and systems that fail completely the second a customer with a thick accent tries to interrupt.

Building truly conversational voice AI isn't just a model problem; it is a deeply complex, unglamorous systems problem.

SigmaMind AI is quietly building the definitive orchestration layer to solve exactly this.

Founded by Ashish Agarwal and Pratik Mundra, SigmaMind is an enterprise-grade platform that lets companies deploy production-ready voice and chat agents that don't just listen, but actually reason, trigger APIs, and complete tasks in real time.

The true moat here is sheer, unadulterated speed and infrastructural rigor. The team have engineered a system that maintains sub-one-second voice-to-voice latency, even when absorbing the brutal drag of traditional telephony overhead. By processing streaming transcripts and acting before a sentence is even finished, their agents can handle interruptions flawlessly and trigger workflows mid-utterance. They handle the incredibly messy audio plumbing so businesses can focus purely on logic and behavior.

The market? Customer support teams, call centers, and global e-commerce brands who are already routing over a million calls a month through the platform. They realize that in the world of voice AI, a one-second delay is the difference between a resolved ticket and a lost customer.

What stands out here is the audacity of the build. It takes immense technical courage to look at a highly fragmented ecosystem and decide to build the robust, connective tissue from the ground up. This is an Indian team building the complex cognitive infrastructure for global conversational AI.

This is the kind of rigorous, head-down execution that gives momentum to the Product Nation dream.

Let's celebrate the builders.

w/ Jay Ingle & Dikshant Joshi

#VoiceAI #ProductNation

[Pratik Mundra @pratmun]

We are solving the biggest problem any voice AI developer today faces: to get stuck in the endless build-test-iterate loop. With this, you can try it now and check out our docs here: https://t.co/NiaOuqq8yM

[Pratik Mundra @pratmun]

We just made voice agents self-improving.

Here's how it works 👇

Define your test cases. Run simulated calls. Get structured results back. Feed them into your prompt -- automatically, using Claude, Cursor, or Codex through the SigmaMind API or MCP.

Your voice agents just keep getting better.

[Pratik Mundra @pratmun]

@agustincto Would love to!

[Pratik Mundra @pratmun]

We launched the SigmaMind MCP server on Product Hunt today.

5M+ calls processed. 2,500+ voice agents in production. And now you can build and manage all of it from inside Claude -- no dashboard, no boilerplate.

Upvote/comment if you want to support 👇
https://t.co/PJPz6IPu7r

[Pratik Mundra @pratmun]

@garrytan We built MCP for voice agents. Would appreciate some feedback - https://t.co/pi3xyoehrc

[Pratik Mundra @pratmun]

@viraj_sheth Flew 2 days ago via Abu Dhabi. Nothing to worry about. Was smooth.

[SigmaMind AI @sigmamind_ai]

You know you want to build a #VoiceAI Agent? Just get on #Cursor or #ClaudeCode and describe it!

One instruction. And then it's done. Watch how!

#Model picked, #Voice configured, #prompt written, phone number assigned. The agent is ready to take calls!

#MCP #Codex #VSCode

[Pratik Mundra @pratmun]

Whoever from @IndiGo6E marketing team thought playing an annoying loud jingle on loop is a good idea should be fired! Just stopping this jingle would go a long way towards improving customer experience. Why even play it?

[SigmaMind AI @sigmamind_ai]

MCP in action: this demo shows how you can run a complete #VoiceAI workflow directly from your #coding environment.

Buy a number → assign an agent → make a call → fetch logs

All via simple prompts to an AI assistant.

Powered by SigmaMind #MCP.

https://t.co/5Y2rFa4OrV

[Pratik Mundra @pratmun]

@ApplyWiseAi @sigmamind_ai Coz we are a smoooth operatorrr… :)

[Pratik Mundra @pratmun]

Launched our MCP server!

You can now build voice agents from within your favorite coding tools using @sigmamind_ai MCP.

MCP makes it easy for anyone to build and launch their voice agents using AI tools.

You can continue to use our APIs if thats your game!

[Pratik Mundra @pratmun]

Get your SigmaMind API key and try it out here: https://t.co/9104QHk3OX

[Pratik Mundra @pratmun]

@adisingh Congrats Adi!

[Pratik Mundra @pratmun]

@arushi_ressl great read!

[Pratik Mundra @pratmun]

@arushi_ressl 💯