SayItDev: Run Apple Intelligence Locally on macOS
SayItDev is a lightweight command-line interface and local server that exposes Apple Intelligence capabilities. It provides fully local text-to-speech, transcription, and an OpenAI-compatible endpoint without requiring cloud APIs or API keys.
Impact: Medium
Why it matters
It allows developers to utilize Apple's native on-device AI capabilities and audio features directly through a CLI or local server, mimicking OpenAI's API locally without sending data to the cloud.
TL;DR
- 01SayItDev runs Apple Intelligence locally on macOS, providing voice and LLM features with no cloud or API keys.
- 02It serves an OpenAI-compatible API locally, exposing endpoints like chat completions, speech synthesis, and transcription.
- 03It requires Microphone and Speech Recognition permissions in macOS System Settings for voice-based features.
Key facts
- Platform
- macOS
- API Compatibility
- OpenAI-compatible endpoints
- License
- Open Source
Native CLI and Local Server
SayItDev operates entirely on-device, offering a set of command-line flags to handle various audio and LLM tasks. Running --transcribe performs speech-to-text conversion locally, which can be further customized with --locale en-US and --timestamps flags to output precise time ranges for each segment.
OpenAI-Compatible API Endpoints
By spinning up the local server, developers get access to multiple standardized endpoints:
/v1/chat/completionsfor local text generation./v1/audio/speechfor text-to-speech synthesis./v1/audio/transcriptionsfor speech-to-text conversions.
This makes it drop-in compatible with existing SDKs and tools designed for OpenAI's cloud API.
Installation and Permissions
To build SayItDev from source, clone the repository and run:
make build && sudo make installFor features requiring microphone access (such as --listen), you must grant Speech Recognition and Microphone permissions to your terminal application within System Settings → Privacy.
Try it in 2 minutes
make build && sudo make installbash
✓ When to use
- When you want to run Apple Intelligence and voice tasks locally on macOS.
- When you need an OpenAI-compatible local API for chat, speech, or transcribing audio on Mac.
✕ When NOT to use
- When you are not on macOS or do not have Apple Intelligence features available.
- When you require enterprise-grade cloud LLMs with high parameter counts.
What to do today
- Clone the repository and run `make build && sudo make install` to build SayItDev.
- Grant terminal permissions for Microphone and Speech Recognition in macOS System Settings to use `--listen` or `--transcribe`.
Sources