https://github.com/openai/openai-realtime-embedded-sdk
Instructions on how to use the Realtime API on Microcontrollers and Embedded Platforms
https://github.com/openai/openai-realtime-embedded-sdk
Last synced: 26 days ago
JSON representation
Instructions on how to use the Realtime API on Microcontrollers and Embedded Platforms
- Host: GitHub
- URL: https://github.com/openai/openai-realtime-embedded-sdk
- Owner: openai
- License: mit
- Created: 2024-11-14T20:44:53.000Z (5 months ago)
- Default Branch: main
- Last Pushed: 2025-03-05T16:21:17.000Z (about 1 month ago)
- Last Synced: 2025-03-12T23:00:41.425Z (about 1 month ago)
- Language: C++
- Homepage:
- Size: 7.81 KB
- Stars: 1,496
- Watchers: 119
- Forks: 169
- Open Issues: 23
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
- awesome-ChatGPT-repositories - openai-realtime-embedded-sdk - A SDK to using the Realtime API with Microcontrollers like the ESP32 (Openai)
README
# Open RealtimeAPI Embedded SDK
# Table of Contents
- [Docs](#docs)
- [Installation](#installation)
- [Usage](#usage)
## Platform/Device Support
This SDK has been developed tested on a `esp32s3` and `linux`. You don't need any physical hardware
to run this SDK. You can use it from Linux directly.
To use it on hardware purchase either of these microcontrollers. Others may work, but this is what
has been developed against.
* [Freenove ESP32-S3-WROOM](https://www.amazon.com/gp/product/B0BMQ8F7FN)
* [Sonatino - ESP32-S3 Audio Development Board](https://www.amazon.com/gp/product/B0BVY8RJNP)
You can get a ESP32S3 for much less money on eBay/AliExpress.
## Installation
Call `set-target` with the platform you are targetting. Today only `linux` and `esp32s3` are supported.
* `idf.py set-target esp32s3`
Configure device specific settings. None needed at this time
* `idf.py menuconfig`
Set your Wifi SSID + Password as env variables
* `export WIFI_SSID=foo`
* `export WIFI_PASSWORD=bar`
* `export OPENAI_API_KEY=bing`
Build
* `idf.py build`
If you built for `esp32s3` run the following to flash to the device
* `sudo -E idf.py flash`
If you built for `linux` you can run the binary directly
* `./build/src.elf`
See [build.yaml](.github/workflows/build.yaml) for a Docker command to do this all in one step.
## Usage