flow.injectAudio
Introduction
flow.injectAudio is a Flow Command that prepares a prerecorded audio file for microphone-based testing.
It is designed for applications that rely on voice input, such as speech-to-text systems, AI assistants, voice search, WebRTC applications, and other microphone-driven experiences.
Rather than requiring a physical microphone during automation, Flowstride works with Chromium's fake media stream capabilities to provide a virtual microphone backed by a prerecorded audio file.
The Problem It Solves
Testing voice-driven applications, such as AI voice assistants, customer service chatbots, voice search, or WebRTC applications, is one of the most challenging areas of UI automation.
When a web application requests microphone access, browsers normally expect input from a real hardware microphone.
Automation frameworks cannot physically speak into a microphone, making these workflows difficult to automate reliably.
flow.injectAudio provides a consistent way to prepare prerecorded audio for automated voice testing.
How flow.injectAudio Works
Flowstride uses Chromium's built-in fake media stream capabilities to provide virtual microphone input.
However, because of how Chromium is architected, there is an important implementation detail that every Flowstride user should understand.
Browser Launch Spoofing
During browser initialization, Flowstride launches Chromium with media stream flags that:
- Automatically grant microphone permission.
- Create a virtual microphone device.
- Configure Chromium to read audio from a prerecorded
.wavfile.
This allows voice-enabled applications to receive audio without requiring any physical recording hardware.
Audio Validation
When flow.injectAudio executes, Flowstride safely resolves the supplied file path relative to the current working directory.
The framework then verifies that the audio file exists before continuing.
If the file cannot be found, execution immediately stops with a descriptive error, preventing silent failures later in the Scenario.
Execution Telemetry
After validation succeeds, Flowstride records the active audio source for the current execution.
This provides clear execution telemetry and confirms which prerecorded audio asset is being used during the test run.
Current Implementation
Chromium requires fake microphone audio to be configured before the browser launches.
Because of this browser limitation, the current implementation of flow.injectAudio primarily serves as a validation and telemetry command during Scenario execution.
The actual virtual microphone is configured during browser initialization.
Syntax
flow.injectAudio "<audio-file>";Parameters
| Parameter | Required | Description |
|---|---|---|
| Audio File | ✅ | Path to the prerecorded audio file. Audio assets should be stored inside the project's fixtures directory. |
Examples
Prepare a voice recording
Feature: Voice Search
Scenario: Search using voice input
Given "Initialize the virtual microphone"
flow.injectAudio "fixtures/customer-question.wav";
When "Open the voice search page"
flow.open "/voice-search";
When "Start voice recording"
flow.click button "Start Recording";
When "Wait for the voice processing pipeline"
flow.waitForPipeline;
Then "The spoken request is transcribed"
flow.expect transcript contains "find nearby coffee shops";Prepare an audio file for voice authentication
Feature: Voice Authentication
Scenario: Authenticate using voice
Given "Load the authentication sample"
flow.injectAudio "fixtures/voice-login.wav";
When "Open the authentication page"
flow.open "/voice-auth";
When "Begin voice verification"
flow.click button "Verify Voice";
When "Wait for processing"
flow.waitForPipeline;
Then "Authentication succeeds"
flow.expect visible "Authentication Successful";When to Use flow.injectAudio
Use flow.injectAudio whenever a Scenario requires microphone input from a prerecorded audio file.
Typical examples include:
- Voice authentication
- Speech-to-text applications
- AI voice assistants
- Voice search
- Customer support bots
- WebRTC voice workflows
Important Notes
Initialize the virtual microphone early
flow.injectAudio should be one of the first commands executed in a Scenario.
This ensures the audio asset is validated before the application begins requesting microphone input.
Store audio files in the fixtures directory
By convention, prerecorded audio assets should be stored inside your project's fixtures directory.
flow.injectAudio "fixtures/customer-question.wav";This keeps test assets organized and portable across environments.
Current Implementation
The current implementation validates the audio file and records the active audio source during execution.
The virtual microphone itself is configured when Chromium is launched.
Use compatible WAV files
Chromium's fake microphone feature is designed to work with .wav audio files.
For best compatibility, use standard PCM WAV files.
Common Mistakes
Do not initialize the microphone after recording starts
Always prepare the audio before interacting with controls that request microphone access.
Correct:
Given "Initialize the virtual microphone"
flow.injectAudio "fixtures/customer-question.wav";
When "Start recording"
flow.click button "Start Recording";Incorrect:
When "Start recording"
flow.click button "Start Recording";
When "Load the audio"
flow.injectAudio "fixtures/customer-question.wav";Do not reference missing audio files
Flowstride validates the audio file before execution.