unpod logo
AboutBlogDocs
Github
The stack · one platform, three layers
Speech Infra
Developer Platform
Telephony + Speech

The metered speech & telephony your agents run on. STT, TTS, numbers and routing through one webhook.

unpod.dev ↗
Open source
Unpod
Open-source CPaaS

Self-host the full agents & contact-center platform. Embeds SuperDialog, bills against Unpod infra.

unpod.ai/open-source-cpaas ↗
Open source
SuperDialog
Conversation framework

Write realtime conversation logic. Standalone, lower-level, free - adopt it before you ever need the cloud.

unpod.ai/superdialog ↗
→ How it works→ Compare→ Quickstart
Open source · self-host · or run on Unpod
unpod logo

Realtime voice + messaging infrastructure. An open-source conversation framework and CPaaS, on metered telephony + speech - so any agent can call and text the real world.

GitHubDocumentation

Products

  • Developer Platform
  • Unpod (OSS CPaaS)
  • SuperDialog

Developers

  • Documentation
  • GitHub
  • API reference

Company

  • About
  • Privacy
  • Terms
© 2026 Unpod.ai · All rights reservedMade in India · Built for the world
About Unpod

New-age interfaces between humans, agents, and the software they run.

Unpod builds the human-machine interface layer for the agent era - so people and agents can communicate naturally with any agent or application, cut the friction out of existing systems, and turn every interaction into productivity.

Talk to the team →Read the docs
Our mission

Communication between humans and machines is still broken.

For decades, software has been built for humans clicking through forms, dashboards and APIs. Every system speaks its own language, and bridging them means glue code, manual handoffs and lost time.

Now agents are everywhere - but they inherit the same fragmented interfaces. An agent that wants to reach another agent, act inside an existing application, or loop in a human still has no natural way to do it.

Unpod builds the interface layer that closes the gap. Human or agent, on either side - communicating through voice, messaging and structured protocols, so the systems you already run get faster instead of heavier.

We're not adding another app to your stack. We're building the connective tissue that lets everything in it finally talk.

- The Unpod team
The interface layer
Initiates
Humansvoice · chat
Agentsautonomous
intent
▶
result
◀
Acts on
Agentsagent ↔ agent
Applicationsexisting systems
Unpodinterface layer
VoiceMessagingProtocol / MCP
One layer translating between every party - reducing inefficiency in systems you already have.
How we build

Four principles we don’t compromise on.

These show up in the code, the pricing, and how we treat the people and agents that build on us.

Interfaces over integrations

We replace brittle glue code with natural interfaces - voice, messaging and structured protocols any human or agent can use.

No glue code

Agent-native by design

Built for agent-to-agent and human-to-agent traffic. Native MCP means Claude Code, Cursor, Windsurf and Zed plug in on day one.

Native MCP

Fits your existing systems

No rip-and-replace. Unpod layers onto the applications and workflows you already run - cutting friction instead of adding a tool.

Drop-in

Open & compliance-grade

MIT-licensed CPaaS and an Apache 2.0 conversation framework, fully self-hostable. ISO 27001 / 42001 certified, DPDP-compliant, data pinned to your region.

Open source
The team

Small team, big ambition.

Building the interface layer takes a rare mix of disciplines. These are the ones that matter most - and the core of each already sits in-house.

✓ On the team

Realtime infrastructure

Sub-second voice and protocol pipelines. In this space the latency budget is the product.

✓ On the team

Conversation & agent design

Shaping messy human intent into reliable behaviour across voice, text and structured protocol.

✓ On the team

Applied speech & AI

STT, TTS and routing tuned for naturalness at scale - the difference between a demo and a deployment.

✓ On the team

Developer experience

SDKs, MCP and docs clean enough that a human or an agent ships in minutes, not weeks.

Hiring

Distributed systems & telephony

Carrier-grade routing, failover and uptime that holds when millions of interactions hit at once.

Hiring

Trust & compliance

Security, data residency and certifications, so the interface layer is safe to put in front of real systems.

5M+
interactions handled
across humans & agents
< 330ms
first-response latency
voice + protocol
99.95%
uptime
failover built-in
Apache 2.0
open source
framework + CPaaS

Want to build the interface layer with us?

We’re hiring engineers, talking to design partners and shipping fast. If human-machine interfaces for the agent era is the thing you want to spend your years on - let’s talk.

Talk to the teamView open roles ↗