Conversational AI platform built for governed scale
Design agents in plain English, generate versioned Agent Spec JSON, test in the browser, then deploy across hundreds of workspaces — every tool call validated, permissioned and logged.
Spec-first, not prompt-only
Describe the caller experience; VoiceForge generates Zod-validated Agent Spec JSON with intents, tools, confirmation rules and handoff. Versioned and testable before any telephony.
Multi-tenant by design
One workspace per client or location: isolated calendars, knowledge, numbers, transcripts and billing. No cross-tenant leakage — the contract agencies need to resell with confidence.
Compliance in the path
Every outbound call passes the gate: consent, opt-out, calling window, AI disclosure and recording notice. Cold outreach blocked by default — governance is operational, not a docs footnote.
One runtime, one spec
Browser tests and phone calls use ElevenLabs for STT, conversation reasoning, and TTS. The same Agent Spec governs every channel through the provider adapter.
From pilot to fleet
Browser test calls with full transcripts, then publish the version you tested. Analytics per workspace — booked, transferred, opted out — turn delivery into retention and upsell.
Questions, answered.
What is a conversational AI platform?
A platform to design, test, deploy and monitor AI voice agents at scale — spec-first Agent Spec JSON, versioned, workspace-scoped, with compliance gates and audit logs. VoiceForge runs browser and phone calls on ElevenLabs.
How does VoiceForge differ from Vapi or Retell for enterprise?
Vapi and Retell are developer APIs. VoiceForge adds governed delivery: declarative specs, Zod-validated tools, per-workspace isolation, consent and window checks that block cold outreach by default, and white-label for agency resale.
Is the platform SOC 2 and HIPAA ready?
Workspace isolation, audit logging, permissioned tool calls and retention controls are built in. BAA and data-retention configuration are available on enterprise plans — ask for the compliance pack and deployment architecture.
Can it scale across hundreds of sub-accounts?
Yes. Each client or location is an isolated workspace with its own calendars, knowledge base, numbers and branding. Flat $49-$499 plans keep unit economics predictable at scale.
Hear the difference.
Then build your own.
Start with a free workspace. Review your agent. Test before you publish.