Guava

Founded 2012

Deploy compliant voice agents for phone calls using Python and TypeScript SDKs

Pricing model

Free Tier, Paid

You can use Guava to build and deploy compliant conversational voice agents directly through code instead of prompt engineering. The system provides Python and TypeScript SDKs alongside a command-line interface to set up inbound and outbound phone workflows. It includes a proprietary speech recognition and synthesis stack that handles turn-taking, speech interruptions, and call audio natively. Guava connects to your existing backend databases and business logic while maintaining regulatory standards like HIPAA, SOC 2 Type II, and PCI DSS Level 1.

What Guava does

Scaffold inbound and outbound voice agent projects using the CLITest telephone conversation flows locally before shipping to productionDefine strict compliance boundaries and business rules directly in codeConnect call logic to external databases and backend APIsProcess spoken card details securely following PCI DSS standardsDeploy voice telephony applications across global telephone networksCode-based conversation rules versioned and tested without prompt driftIntegrated speech recognition and audio synthesis stack without third-party dependenciesNatural turn-taking engine that manages live interruptions and pausesCLI tool for scaffolding, testing locally, and deploying telephone agentsDirect integration with enterprise databases without data duplicationBuilt-in compliance coverage including SOC 2 Type II, HITRUST i1, and PCI DSS Level 1Multi-tenant and multi-language support with full audit trails on calls