# Hamming

> Automated testing and production monitoring platform for voice and chat AI agents with 50+ built-in metrics and compliance support.

Hamming is the category-leading platform for voice agent quality assurance, providing automated testing and production monitoring for AI-powered voice and chat agents. Built by a team with experience scaling ML systems at Tesla and mission-critical infrastructure at Citizen, Hamming helps teams catch issues before they impact customers through comprehensive pre-launch testing and continuous production monitoring.

The platform enables teams to validate agent behavior across diverse scenarios including accents, background noise, interruptions, and edge cases. With support for 65+ languages and regional accents, Hamming simulates real-world conditions to ensure voice agents perform reliably in production environments.

- **Auto-Generated Test Scenarios** - AI automatically generates hundreds of test cases from your agent's prompt, covering happy paths, edge cases, and adversarial inputs without manual test case writing
- **Production Call Monitoring** - Score 100% of production calls automatically with real-time quality scoring, detecting hallucinations, policy violations, and quality degradation as they happen
- **50+ Built-in Metrics** - Track latency, hallucinations, sentiment, compliance, turn-taking, and more with detailed component-level breakdowns for STT, LLM, and TTS
- **Load Testing at Scale** - Run 1,000+ concurrent test calls to stress-test infrastructure and identify breaking points before peak traffic impacts customers
- **CI/CD Integration** - Trigger tests on every deploy via REST APIs, integrate with GitHub Actions or Jenkins, and block bad prompts from reaching production
- **Production Call Replay** - Convert any failed production call into a replayable test case with one click, preserving original audio, timing, and caller behavior
- **Compliance and Security** - SOC 2 Type II certified with HIPAA BAA available, featuring red-teaming for prompt injection, jailbreaks, and PII leakage testing
- **Multi-Platform Support** - Native integrations with Vapi, Retell, 11Labs, LiveKit, Pipecat, and more with SIP and WebRTC connectivity options

To get started, connect your agent by dialing a SIP number or via WebRTC, paste your prompt or connect via API, and run your first test in under 10 minutes. Hamming auto-generates test suites and provides detailed reports with call recordings, transcripts, and metrics for comprehensive voice agent QA.

## Features
- Auto-generated test scenarios from agent prompts
- Production call monitoring and analytics
- 50+ built-in evaluation metrics
- 1000+ concurrent call load testing
- 65+ language and accent support
- CI/CD pipeline integration
- Production call replay and regression testing
- Red-teaming for security vulnerabilities
- Custom evaluation metrics
- Real-time alerts via Slack and webhooks
- DTMF and IVR emulation
- Detailed PDF test reports
- Speech-level analysis beyond transcripts
- OpenTelemetry observability integration
- Role-based access control

## Integrations
Vapi, Retell AI, ElevenLabs, LiveKit, Pipecat, OpenAI, Synthflow, Daily, GitHub Actions, Jenkins, Slack, PagerDuty, Okta, Azure AD, Google Workspace

## Platforms
LINUX, WEB, API

## Pricing
Subscription-based

## Links
- Website: https://hamming.ai
- Documentation: https://docs.hamming.ai/
- EveryDev.ai: https://www.everydev.ai/tools/hamming