InputRight is an open-source platform designed to deploy AI voice agents that capture conversations and transform them into accurate, user-verified structured data. This repository contains a complete, production-ready system that allows the deployment of voice-enabled AI receptionists capable of answering inquiries and reliably gathering information from users.

Key Features

• Voice-First Interaction: Facilitates natural, hands-free conversation as the main user interface.
• Guaranteed Data Accuracy: Utilizes a unique RPC-driven workflow that enables users to visually verify all captured information in a form, effectively eliminating errors.
• Extensible & Configurable: Users can easily configure the agent’s personality, knowledge base, and operational tools.
• Powered by Professional Technologies: Built on robust technologies including LiveKit for real-time communication, and a solid foundation using Python/FastAPI and React/Next.js.

The Open Core Model

InputRight adheres to a modern open-core model, offering two editions:

• InputRight (Open Source): A free, self-hostable core engine ideal for developers aiming to integrate a powerful voice-to-form agent into their applications.
• InputRight (Cloud) - Coming Soon: A fully-managed, multi-tenant SaaS platform that enables effortless deployment of a "Digital Receptionist" for businesses, requiring zero coding and minimal effort.

Roadmap: The Verified Data Platform

The mission is to deliver the most reliable method of converting conversations into highly accurate, structured data. Future features will enhance the unique Human-in-the-Loop Verification process, allowing high-value actions driven by verified data.

Leveraging Verified Data with Advanced Tools

Once data is confirmed by the user, it becomes a trusted asset. Planned tools include:

• Appointment Booking: Use verified details to book appointments in calendars (e.g., Google Calendar, Calendly) with zero risk of typographical errors.
• CRM Integration: Seamlessly create and update customer records in platforms like HubSpot or Salesforce using perfectly accurate data.
• Custom Workflows: Trigger internal processes via webhooks utilizing verified data sent to invoicing systems, project management tools, or custom software.

Enhancing Verification with Multimodal Input

Expanding workflows to incorporate multiple input methods for capturing richer and more accurate data:

• Live Video Analysis: Users can show problems via webcam, facilitating real-time visual confirmation of issues by the agent.
• File & Document Uploads: Users can upload photos or documents for processing, allowing the agent to extract and verify relevant information with the user.

Building Trust with AI Avatars

Enhancements will be made to provide engaging and trustworthy interactions by integrating with leading real-time avatar providers, giving agents a more human-like presence to foster connection and improve user engagement.

Security and Compliance Disclaimer

The open-source version is provided as-is and may not comply with regulations such as HIPAA or PCI DSS out of the box. Ensuring that implementations meet legal and regulatory requirements is the responsibility of the user. The upcoming InputRight Cloud service will offer enterprise-grade security and compliance features tailored for regulated industries.