Update documentation

[MikeBirdTech]

[WIP]

A re-structure of the docs

Easiest way to review is to fetch and check out the branch, cd docs, then run mintlify dev to see a preview in your browser

[benxu3]

we moved the flags and settings mentioned in the docs here to be set in a profile under the software/source/server/profiles directory. this was changed as part of the 01 refactor with @KillianLucas, it would be great if you could reflect these changes by modifying this doc.

i.py, --llm-service, --tts-service, and --stt-service flags are all deprecated now. we only support faster_whisper for stt now as a result of the refactor to use Kolja's RealtimeSTT and TTS. the previous cli flag settings are now configured in a profile.

for example rather than specifying a --tts-service through the flag, you would set interpreter.tts = 'openai' programmatically in a profile and then import the interpreter from that profile in async_server.py. see the current async_server.py for an example of this!

[MikeBirdTech]

Thanks!

Does that include --expose?

Are there any CLI flags that are still valid?

[KillianLucas]

Excellent work Mike. This is such an improvement, I think people will find this way easier to understand. My thoughts:

I think the Using 01 section can be removed, and both articles under that can be pushed into the Configure article of Software Setup — just showing how, in a profile, you can change the LLM (same as in OI) and the tts with interpreter.tts.

For Hardware, I think breaking it down under the 01 Light (with sub-articles like: Materials, Chip (ESP32), Case, Assembly — rough ideas) would be helpful. Ideally embedding James' video somewhere, and Thomas Smith's video about Flashing the 01 Light (maybe in anIntro article, or in their respective sub-articles). No need to have the M5 Atom be a separate article I think.

A few things I think could belong in the Hardware section as well:

• Custom Hardware with like a "You can build your own custom hardware that uses the 01 server. Simply start the server [linking to those docs], send LMC audio chunks to the websocket at / ..." That kind of thing, just explaining the client/server abstractly.
• Desktop (basically a duplicate of the Server / Run article)
• iOS/Android, which links to 1) our react-native WIP app, and mostly is similar to its README that explains how to use it with Expo, and 2) the community native iOS app by eladekkal.

Let me know what you think!

[KillianLucas]

Nice, thanks for this update Mike. Merging now, will push to Mintlify.