Coming soon · Early access for design partners

The platform powering personal AI assistants for business applications.

MoltiCast and MoltiEdge let any consumer app deploy personalized, domain-specific agents at scale — turning AI into a core product layer, not a side feature.

Request early access See how it works

embed-agent.ts

import { Vuel } from "@vuel/sdk";

const vuel = new Vuel({ apiKey });

// give every user their own agent
const agent = await vuel.agents.create({
  userId:      "user_8f3k2",
  skills:      ["calendar", "calls", "receipts"],
  personality: "warm-professional"
});

Embed in days

Drop into any consumer app

Always-on

Persistent, not stateless

The problem

Every app wants an agent. The math doesn't work.

Demand is everywhere — every consumer business wants an embedded assistant inside its product. The infrastructure to actually ship one isn't.

Demand is everywhere

Fitness, banking, travel, retail, healthcare — every consumer business wants an embedded AI assistant inside its product.

ii.

Today's stack doesn't fit

Existing AI runtimes were built to host one agent at a time. Giving every user their own assistant becomes prohibitively expensive — fast.

iii.

So nothing ships

Teams either build something shallow they don't trust, or they punt the roadmap. The user never gets the agent.

01.

MoltiCast — the runtime.

A multi-tenant agent runtime built ground-up for scale. Every tenant ships their own agent — same runtime, custom skills, isolated memory, branded personality.

Always-on. Persistent agent, not a stateless API call
Skill-driven. Tenant-controlled skill graph
Tenant-isolated. User data and personality stay bound
Embed in days. Drop into any consumer app via SDK

MoltiCast · agent spec per tenant

Per-tenant agent

Configured

Identity

personabranded · custom

voiceyour tone

nameyour-agent

Skills · tenant-controlled

Calendar Voice calls Inbox triage Receipts Files + custom

Isolation

memoryper-user

dataencrypted at rest

modelmanaged

02.

MoltiEdge — the services.

A headless service layer agents call directly. Headless web, voice calls, and a growing catalog of business integrations — pay-as-you-go, plug into MoltiCast or your own stack.

Browse, fill, scrape, transact on the open web
Place real outbound voice calls, end-to-end
Slack, Notion, JIRA, Zoho and more
Bring your own agent — fully headless API

MoltiEdge · services 4 active

Headless web

Browse, fill, scrape, transact

Voice calls

Outbound calls, end-to-end

Third-party tools

Slack, Notion, JIRA, Zoho

Bring your own agent

Headless. Plug into any stack

Developer SDKs

Drop into any app. In days.

Two SDKs cover the integration surface — server-side hooks for MoltiEdge services, and a mobile UI kit for the agent's face inside your app. Both work with MoltiCast or your own agent stack.

Server SDK · MoltiEdge API

vuel-edge

Your agents act on the world.

Call MoltiEdge services — headless web, voice calls, third-party tools — directly from your server code with type-safe clients and idiomatic patterns for each language.

REST and gRPC clients, auto-generated types
Webhooks for async results & long-running tasks
Built-in retries, observability, and tracing
Plug into MoltiCast or your own agent stack

import { Vuel } from "@vuel/edge";

const vuel = new Vuel({ apiKey });

// place an outbound voice call
await vuel.voice.call({
  to:     "+15552480193",
  intent: "check refill status",
  onComplete: "webhook://refills"
});

Node Python Java Go

Mobile SDK · Agent UI

vuel-mobile

Your agent's face, ready to ship.

A drop-in UI kit for the agent inside your iOS or Android app — chat surface, voice input, briefing cards, tool browser — themeable to match your brand.

Chat, briefing, calls, and tool surfaces out of the box
Voice input, push notifications, background tasks
Themeable — colors, fonts, bubbles, layout
Works with MoltiCast or any agent backend

// SwiftUI · drop the agent surface anywhere
import VuelMobile

VuelChatView(
    agentId: "agent_8f3k2",
    theme:   .brand(.midnight)
)
.vuelOnAction { action in
    // hand off to your code when needed
}

iOS · Swift Android · Kotlin

Open by design OpenAPI specs, MIT-licensed clients, semantic versioning. We won't lock you in.

Request SDK access →

Plans

Two ways to embed. Designed around your stack.

We're working with design partners on pricing tailored to your scale and requirements. Reach out and we'll match you to the right plan.

Business

Shared infrastructure.

For consumer apps embedding agents

Talk to us for pricing.

Multi-tenant runtime
Vuel-managed model selection
User management dashboard
MoltiEdge included
Standard SLA

Request Business →

Enterprise

Dedicated infrastructure.

For regulated and high-scale customers

Custom pricing.

Dedicated runtime
Customer-selectable models
User management dashboard · SSO
MoltiEdge included
Hardened security

Request Enterprise →