FluidVoice System Requirements: macOS Version, Hardware, and Disk Space Guide

FluidVoice requires macOS 15.0 (Sequoia) or later, an Apple Silicon Mac for most AI models (Intel Macs supported with Whisper only), ~1–3.5 GB of disk space, and microphone and accessibility permissions.

FluidVoice is an open-source macOS dictation application by altic-dev that runs AI speech-recognition models entirely on-device or with optional cloud providers. Before installing, verify that your Mac meets the FluidVoice system requirements detailed below, as documented in the repository's README and enforced in the source code.

macOS Version and Hardware Requirements

Minimum Operating System

The app targets macOS 15.0 (Sequoia) as its baseline. In Sources/Fluid/fluidApp.swift, the entry point explicitly validates this:

guard ProcessInfo.processInfo.isOperatingSystemAtLeast(
        OperatingSystemVersion(majorVersion: 15, minorVersion: 0, patchVersion: 0))
else {
    fatalError("FluidVoice requires macOS 15.0 (Sequoia) or later.")
}

This check occurs at app launch, immediately terminating if the requirement isn't met. The deployment target is also declared in Package.swift as .macOS(.v15).

Apple Silicon vs. Intel Macs

• Apple Silicon Macs – Required for all bundled speech models including Nemotron, Parakeet, and other on-device AI models.
• Intel Macs – Supported only when using Whisper models, available from FluidVoice version 1.5.1 onward.

This distinction exists because the bundled AI models leverage Apple Silicon's Neural Engine for acceleration, while Whisper provides a fallback path via CPU inference on Intel hardware.

Disk Space Requirements

Model storage demands vary by capability:

Model Type	Disk Space	Purpose
Voice model	~1 GB	Core speech-to-text recognition
Fluid Intelligence	~3.5 GB	Optional local AI enhancement for post-processing and intelligent formatting

Both models download on-demand after installation. Plan for up to 4.5 GB total if enabling all features.

Required macOS Permissions

FluidVoice needs two system permissions to function, as declared in Info.plist and requested at runtime:

Microphone Access

Required for audio capture. macOS presents a permission dialog automatically when the app first accesses audio input.

Accessibility Permissions

Required for the app to type transcribed text into other applications. The initialization code in fluidApp.swift triggers this prompt:

private func requestAccessibilityPermission() {
    let options: NSDictionary = [kAXTrustedCheckOptionPrompt.takeRetainedValue() as String: true]
    let _ = AXIsProcessTrustedWithOptions(options)
}

Without this permission, FluidVoice cannot insert text into external apps, rendering its core dictation workflow non-functional.

How the App Enforces Requirements at Launch

The complete initialization flow in Sources/Fluid/fluidApp.swift demonstrates how FluidVoice validates its environment:

import SwiftUI
import AppKit

@main
struct FluidApp: App {
    init() {
        // Verify macOS version (must be macOS 15.0 or newer)
        guard ProcessInfo.processInfo.isOperatingSystemAtLeast(
                OperatingSystemVersion(majorVersion: 15, minorVersion: 0, patchVersion: 0))
        else {
            fatalError("FluidVoice requires macOS 15.0 (Sequoia) or later.")
        }
    }

    var body: some Scene {
        WindowGroup {
            ContentView()
                .onAppear {
                    // Prompt the user for microphone and accessibility access
                    requestMicrophonePermission()
                    requestAccessibilityPermission()
                }
        }
    }

    private func requestMicrophonePermission() {
        // Triggers macOS permission dialog on first audio access
    }

    private func requestAccessibilityPermission() {
        let options: NSDictionary = [kAXTrustedCheckOptionPrompt.takeRetainedValue() as String: true]
        let _ = AXIsProcessTrustedWithOptions(options)
    }
}

This mirrors the real implementation at lines 12–24 of Sources/Fluid/fluidApp.swift.

Key Source Files Defining Requirements

File	Role
README.md	Official system requirements documentation
Sources/Fluid/fluidApp.swift	App entry point; enforces OS version and requests permissions
Package.swift	Swift Package Manager manifest with .macOS(.v15) deployment target
Info.plist	Privacy usage descriptions for microphone and accessibility
Sources/CoreAudioCaptureSupport/	Native audio capture requiring macOS audio APIs

Summary

• FluidVoice system requirements center on modern macOS hardware and software
• macOS 15.0 (Sequoia) minimum, hard-enforced at launch
• Apple Silicon required for full AI model support; Intel limited to Whisper
• 1 GB minimum disk space, 4.5 GB recommended for all features
• Microphone and accessibility permissions mandatory for core functionality
• All requirements documented in README.md and enforced in fluidApp.swift

Frequently Asked Questions

Can I run FluidVoice on macOS 14 or earlier?

No. The app explicitly checks for macOS 15.0 at launch and terminates with a fatal error on older versions. This requirement is non-negotiable due to API dependencies in the audio capture and AI model layers.

Does FluidVoice work on Intel Macs?

Limited support exists. Intel Macs can only use Whisper models starting from version 1.5.1. The bundled Nemotron, Parakeet, and Fluid Intelligence models require Apple Silicon's Neural Engine and will not load on Intel hardware.

How much disk space should I allocate for FluidVoice?

Reserve 1 GB minimum for basic voice recognition, or 4.5 GB if you plan to download the optional Fluid Intelligence model. Models download after installation, so initial app size is smaller than the full working footprint.

Why does FluidVoice need accessibility permissions?

The accessibility permission allows FluidVoice to programmatically type transcribed text into other applications. Without it, the app can capture and process audio but cannot output results to your active text field, breaking the core dictation workflow.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}