Intervo: Build Voice and Chat AI Agents

Overview

Intervo is a web platform for building and managing AI agents that handle voice calls and chat. Its visual workflow editor lets teams create multi-step conversations, route users by intent, and coordinate specialized agents to complete tasks.

It is aimed at teams that want to deploy conversational agents using their own infrastructure and connect them to business data or phone systems. The platform brings together a frontend, backend, and RAG service, with integrations for speech providers, LLMs, Twilio, and Stripe.

Key Features

• Build visual, node-based conversation workflows and route conversations based on intent.
• Set goals for specialized sub-agents, including required tasks such as collecting a user's email.
• Create agents for real-time voice calls and text-based web chat.
• Index files, website content, text, and FAQs for retrieval-augmented generation using ChromaDB.
• Connect phone workflows to Twilio and trigger outbound calls through a REST API.
• Choose among supported speech-to-text, text-to-speech, and LLM providers.
• Embed a customizable React chat widget on a website.
• Use Stripe integrations for billing and subscriptions.

Use Cases

• A customer-support team can build a phone or web agent that answers questions from its knowledge base and routes users through support workflows.
• A sales team can use goal-oriented conversations to qualify leads and collect contact details before follow-up.
• A business with existing phone operations can connect inbound or outbound agent calls through Twilio.
• A web team can add a conversational assistant to a site using the embeddable widget and its own company content.

Getting Started

The README documents a Docker Compose setup. Clone the repository, configure the backend environment, then start the services:

git clone https://github.com/Intervo/Intervo.git
cd Intervo
cp packages/intervo-backend/.env.example packages/intervo-backend/.env
# Edit packages/intervo-backend/.env, including MONGO_URI and JWT_SECRET
docker-compose up -d

See the README for configuration options and local development instructions.

Alternatives

• MaxKB: MaxKB focuses on knowledge-based agents, workflows, and tools, while Intervo is tailored to voice calls and web chat.
• CyberVerse: CyberVerse focuses on real-time voice agents with persistent character memory and optional digital-human video, rather than visual conversation workflows for calls and chat.

Considerations

The project is marked inactive, with no commits in over a year and no releases in the past year. The measured contributor history is concentrated in one person, and no issues were closed recently. The README also identifies test coverage and setup documentation as areas still being improved, so teams should review the code and deployment process before relying on it in production.

Self-hosting requires configuration for MongoDB and any external services the chosen agent features use, such as LLM, speech, or Twilio credentials. The README describes several integrations, but lists WebRTC, mobile SDKs, advanced analytics, and Kubernetes deployment support as roadmap items rather than current features. The repository is licensed under MIT; review the license terms when distributing or modifying the software.