[All work](/work?type=work)

# ArticulateAI

One of my early voice experiments: record a clip, transcribe it with Whisper, send the transcript to GPT with a prompt, and optionally speak the reply.

[Source code](https://github.com/Slaymish/ArticulateAI)

Independent development. Python, OpenAI and ElevenLabs APIs.

![Original ArticulateAI command menu, with twelve options for processing a spoken transcript.](https://cdn.sanity.io/images/qnuj1c4o/production/9e3bdfecff2b2c2c958917824aad775e98a5b9c4-1600x1000.png?w=1280\&q=80\&auto=format)

Twelve starting prompts, so the same audio path handles a question, a lecture or a recipe, and the recording length is chosen before you start. The transcript saves with a timestamp, and short replies go to ElevenLabs for speech.

It's a loop over separate API calls, which is what this looked like in 2023. A conversation is a different problem.

## State

The workflow works. The model names are from 2023, so it needs updating before it runs again.

[All work](/work?type=work)