BadloVoice

Free plan Open source AndroidAPIiPhoneLinuxMacSelf-hosted+2

What BadloVoice is

BadloVoice is a free, MIT-licensed realtime AI voice changer for browsers and Android, intended for Discord, games, streams, and calls. Its README says audio stays on the user’s machine rather than being uploaded to the cloud. Simple mode provides voice presets, pitch adjustment, and recording, while Pro mode adds device controls, VU and spectrum displays, latency modes, and an effects rack. Users can load .pth, .pt, or .onnx models with a companion .index file and switch model slots. Listed effects include robot, whisper, telephone, distortion, bitcrush, chorus, echo, reverb, EQ, and compressor. A converter accepts WAV, MP3, FLAC, OGG, and M4A files and outputs processed WAV audio. Routing to Discord and games uses a virtual audio cable; OBS is also mentioned. Desktop setup covers Windows, macOS, and Linux, with Python 3.11+ and Node.js 20+ recommended in the quick start.

Who it suits

BadloVoice suits users who want local voice changing for calls, games, streams, or recordings. It also fits people comfortable building an Android debug APK or setting up desktop dependencies.

What it does well

Free and MIT licensed
Local audio processing
Realtime presets and pitch controls
Many listed audio effects
Supports processed audio export

What to know first

iOS release is not described
Android use requires building a debug APK
Desktop setup has software prerequisites

Plans

Plan	Billed and limits	Price
MIT open-source	Self-hosted software · no paid tier stated · user supplies models and runtime dependencies	Free

As github.com lists them, read 15 Oct 2026. The maker's own page has the final word.

Verdict

BadloVoice offers local realtime voice changing, model loading, effects, and file conversion without a listed cloud upload. Its iOS support is only a setup note, and Android availability is described through a build process.

Facts from BadloVoice's own pages

Android availability: An Android debug APK can be built with the Android SDK and Gradle using the documented npm run apk command. · github.com, 4 Oct 2026
API and extensibility: The repository exposes a WebSocket audio route and includes Python, React, Capacitor and FastAPI code, but no separate public SDK package is documented. · github.com, 15 Oct 2026
Architecture: The system uses a React and AudioWorklet browser or Capacitor WebView, FastAPI, a WebSocket at /ws/audio and a DSP/model pipeline. · github.com, 15 Oct 2026
Audio effects: The DSP rack lists gate, robot, whisper, telephone, distortion, bitcrush, chorus, echo, reverb, EQ, and compressor effects. · github.com, 4 Oct 2026
Backend: The backend directory contains FastAPI, WebSocket processing, model slots, presets and DSP components. · github.com, 15 Oct 2026
Configuration: Runtime settings use the SHIFTBOOTH_ environment prefix, including host, port, model directory, sample rate, chunk size and latency mode. · github.com, 15 Oct 2026
Desktop requirements: The quick start lists Windows, macOS, or Linux, Python 3.11+ recommended, and Node.js 20+. · github.com, 4 Oct 2026
Documentation: Documentation includes a README, a models guide, screenshots and contribution and security files in the repository. · github.com, 15 Oct 2026
DSP effects: The effects rack lists gate, robot, whisper, telephone, distortion, bitcrush, chorus, echo, reverb, EQ and compressor effects. · github.com, 15 Oct 2026
DSP-only operation: Pitch, formant and effects can run without loading a voice checkpoint. · github.com, 15 Oct 2026
F0 methods: The listed F0 choices are PM, Harvest, Crepe, RMVPE and FCPE. · github.com, 15 Oct 2026
File conversion: The converter accepts wav, mp3, flac, ogg, and m4a files and produces processed WAV audio. · github.com, 4 Oct 2026
GitHub topic classification: The BadloVoice repository is tagged on GitHub with React, Android, Python, open-source, Discord, WebSocket, DSP, Capacitor, RVC, ONNX and FastAPI topics. · github.com, 15 Oct 2026
GitHub topic coverage: GitHub's voice-conversion topic page listed 376 public repositories when opened. · github.com, 15 Oct 2026
GPU: An NVIDIA GPU with CUDA is optional for faster RVC or ONNX processing. · github.com, 4 Oct 2026
GPU support: An NVIDIA GPU with CUDA is optional and documented as a way to accelerate RVC and ONNX processing. · github.com, 15 Oct 2026
Installation: Desktop setup uses git clone, a Python virtual environment, backend requirements, Uvicorn and a Vite frontend. · github.com, 15 Oct 2026
Integrations: The routing documentation covers Discord, games and OBS through a virtual audio cable. · github.com, 15 Oct 2026
iOS note: The README says iOS setup requires a Mac and Xcode; it does not describe an iOS release. · github.com, 4 Oct 2026
License: The project is released under the MIT license. · github.com, 15 Oct 2026
Limit: mobile distribution: The documented Android output is a debug APK build, and the repository does not describe a published store release. · github.com, 15 Oct 2026
Limit: model availability: Full RVC conversion requires users to obtain compatible voice models and supporting assets such as HuBERT. · github.com, 15 Oct 2026
Limit: routing complexity: Using the software with Discord or games requires separate virtual-cable installation and audio-device routing. · github.com, 15 Oct 2026
Local processing: The documented pipeline processes microphone audio locally through DSP, RVC or ONNX and outputs to speakers or WAV. · github.com, 15 Oct 2026