echofire 0.1.2

Realistic Voice Agent Testing - Automate end-to-end evaluation of voice agents with human-like audio simulation

EchoFire

Realistic Voice Agent Testing

Automate end-to-end evaluation of voice agents with human-like audio simulation.

flowchart LR
    A[Record Session With Voice Agent] --> B[Write Tests w/ LLM Judge + Built-in Functions]
    B --> C{Parallel Test Execution}
    C -->|Multiple Iterations| D1[Run Test 1]
    C -->|Multiple Iterations| D2[Run Test 2]
    C -->|Multiple Iterations| D3[Run Test ...]
    D1 --> E[Share Results & Collaborate]
    D2 --> E
    D3 --> E
    E --> F[Build Datasets for Fine-Tuning]
    F --> G[Improve Models & Prompts]
    G --> A

Why EchoFire?

• 🚫 Manual testing sucks: Listening to every agent response isn't scalable.
• 🤖 Synthetic TTS isn't real: Simulate actual human speech patterns, background noise, and ASR edge cases.
• 🔥 Test everything: Validate ASR accuracy, intent logic, and agent responses in one flow.