AI agents that actually resolve.
Riyaan agents handle the full conversation — on your website, on the phone, via SMS, and inside physical robots. Same cloud brain. Same guardrails. Every channel.
I found your booking #RF-4821 on the United flight to Seoul. There's a 10:45 AM departure next Tuesday with window seats available. I can rebook you right now and send the updated confirmation to your email — shall I go ahead?
You’ve shipped this before
Fragile agent stack
Your voice agent works in the demo but breaks in production. Background noise, interruptions, edge-case questions — the real world doesn't follow your script.
Agents running unsupervised
Your agent promised a refund it wasn't authorized to give. You found out three days later from a customer complaint. No transcript, no audit trail, no alert.
Months before a single live call
Guardrails, compliance, human handoff, multi-channel — each one is a quarter of engineering. You're building infrastructure instead of serving customers.
One platform. Full control.
Riyaan gives you production-grade agents with the guardrails, observability, and channel coverage built in.
Industrial-grade from day one
Agents that handle background noise, interruptions, and ambiguous requests — not just clean demo scripts. One platform for voice, web, phone, and SMS.
Every conversation, fully auditable
Live transcripts, tool execution logs, guardrail triggers, and outcome tracking. Know exactly what your agent said, did, and whether it resolved the issue.
Live in days, not quarters
Guardrails, human handoff, knowledge base, compliance controls, and analytics — built in. We've seen teams go from first call to production in under two weeks.
Where agents are already working
Real conversations, real resolutions — across industries that don't tolerate mistakes.
Configure
Define what your agent knows, what it can do, and where it stops. Dashboard or CLI.
Connect
Go live on voice, phone, chat, and SMS from a single agent configuration.
Improve
Review real conversations, catch issues early, and push improvements without downtime.
Configure
Define what your agent knows, what it can do, and where it stops. Dashboard or CLI.
Connect
Go live on voice, phone, chat, and SMS from a single agent configuration.
Improve
Review real conversations, catch issues early, and push improvements without downtime.
Where Riyaan agents are already working
From customer support lines to government service robots — the same platform, in production.
E-commerce support that resolves
An AI agent handling billing disputes, returns, and account changes — with human handoff when needed. Full audit trail on every conversation.
Real-time translation at a government office
A service robot deployed at a government office — greeting foreign visitors and providing live translation, powered by the same Riyaan cloud runtime.
A robot dog that listens and acts
A quadruped robot controlled entirely by voice conversation — performing tricks, expressions, and movements, all decided by the cloud AI in real time.
Dental offices, booked by AI
Appointment scheduling, rescheduling, and reminders handled end-to-end by voice — no hold music, no phone tag.
Built for production
One SDK. Full observability. A path from first call to enterprise scale.
Pick Your Provider
Gemini 3, Flash, embedding models
model: gemini-3-flash-preview
+ DeepSeek, Mistral, xAI, Groq, and 15+ more via Pi
Platform Layer
Riyaan Platform
Agent Terminal
$ riyaan agent deploy --provider google
Deploying with Google...
Model: gemini-3-flash-preview
Guardrails: 3 active rules
Tools: check_warranty, book_appointment
Agent live on wss://your-app.livekit.cloud
$ riyaan test call --scenario billing-dispute
Calling +1 (555) 012-3456...
Connected — testing with background noise and interruptions enabled.
_
import { createAgent } from '@riyaan/sdk/agent';
import { createConvexRoomCallbacks } from '@riyaan/sdk';
const agent = await createAgent({ provider: 'google', model: 'gemini-3-flash-preview' });
const callbacks = createConvexRoomCallbacks({ convexUrl, appSlug, getSessionToken });
// same guardrails, compliance rules, and audit trail across every provider
Pay for outcomes, not seats.
You only pay when your agent actually resolves the conversation. Escalations to humans are free.
Starter
Free
Test with real conversations. See what your agent can handle.
- 100 resolved conversations / mo
- 1 agent
- Community support
- Basic analytics
- Pi Agent SDK
Pro
$0.50 / resolution
Production-grade agents. Pay only when they resolve.
- Unlimited agents & conversations
- Tool integration & guardrails
- A/B testing & experiments
- RAG knowledge base
- 22+ providers (Google, Anthropic, OpenAI, ...)
- Custom extensions & skills
- Multi-channel (web, phone, SMS)
Enterprise
Custom
For regulated industries and teams that need custom outcome definitions, SLAs, and compliance controls.
- Everything in Pro
- Custom outcome definitions
- SSO & role-based access
- Dedicated support & SLA
- Compliance & audit controls
- Dedicated runtime & data isolation
- Custom provider integration