This is a web app that gives GPT-4 and GPT-3.5 a voice interface. Anyone can now verbally communicate with ChatGPT.
Both desktop and mobile browsers are supported, enabling users to talk to GPT-4 on the go.
OpenAI Whisper is used for speech-to-text and OpenAI gpt-3.5-turbo/gpt-4 for completion.
To get it running:
git clone git@github.com:lspahija/VoxGPT.gitcd VoxGPTdocker build -t voxgpt .docker run -d -e OPENAI_API_KEY='<YOUR_API_KEY>' -p 8000:80 voxgpt- navigate to
localhost:8000in a modern browser
The AI model defaults to gpt-3.5-turbo but you can adjust this by setting the AI_COMPLETION_MODEL environment variable (e.g. to gpt-4 if your OPENAI_API_KEY has access to it)
A demo is currently deployed at https://ultimate-ai-assistant.up.railway.app/
Ultimate.AI.Assistant.mp4
Output audio speed can be adjusted by setting the AUDIO_SPEED environment variable e.g. setting this to 1.5 will result in audio playing back at 1.5x default speed.
You can configure the language by setting the LANGUAGE environment variable to the corresponding ISO-639-1 code. The default is en. The supported languages are:
- af: Afrikaans
- ar: Arabic
- bg: Bulgarian
- bn: Bengali
- bs: Bosnian
- ca: Catalan
- cs: Czech
- da: Danish
- de: German
- el: Greek
- en: English
- es: Spanish
- et: Estonian
- fi: Finnish
- fr: French
- gu: Gujarati
- hi: Hindi
- hr: Croatian
- hu: Hungarian
- id: Indonesian
- is: Icelandic
- it: Italian
- iw: Hebrew
- ja: Japanese
- jw: Javanese
- km: Khmer
- kn: Kannada
- ko: Korean
- la: Latin
- lv: Latvian
- ml: Malayalam
- mr: Marathi
- ms: Malay
- my: Myanmar (Burmese)
- ne: Nepali
- nl: Dutch
- no: Norwegian
- pl: Polish
- pt: Portuguese
- ro: Romanian
- ru: Russian
- si: Sinhala
- sk: Slovak
- sq: Albanian
- sr: Serbian
- su: Sundanese
- sv: Swedish
- sw: Swahili
- ta: Tamil
- te: Telugu
- th: Thai
- tl: Filipino
- tr: Turkish
- uk: Ukrainian
- ur: Urdu
- vi: Vietnamese
- zh: Chinese (Mandarin)