mirror of
https://github.com/jambonz/talk-to-your-llm.git
synced 2025-12-19 03:57:43 +00:00
README
This commit is contained in:
@@ -12,14 +12,16 @@ WS_PORT=3000 OPENAI_API_KEY=xxxx node app.js
|
||||
|
||||
In the jambonz webapp create an application with url `wss://<your-domain>/llm-voicebot` and route calls to it.
|
||||
|
||||
## Environment variables
|
||||
## Configuration
|
||||
|
||||
|variable|meaning|required|
|
||||
|Environment variable|meaning|required|
|
||||
|---------|------|--------|
|
||||
|OPENAI_API_KEY|Your api key|yes|
|
||||
|OPENAI_MODEL|model to use|no (default: gpt-4-turbo)|
|
||||
|BOT_NAME|Name of Assistant to create|no (default: jambonz-llm-voicebot)|
|
||||
|
||||
If you want to change the system instruction edit [./data/settings.json](./data/settings.json)
|
||||
|
||||
## Limitations
|
||||
|
||||
Currently, there is no support for adding [Tools](https://platform.openai.com/docs/assistants/tools) to the Assistant. User input is transcribed as presented as simple text and the OpenAI assistant is encouraged (via system instructions) to respond with text that is brief and un-annotated.
|
||||
Currently, there is no support for adding [Tools](https://platform.openai.com/docs/assistants/tools) to the Assistant. User input is transcribed and then presented to the OpenAI Assistant as simple text, and the OpenAI assistant is encouraged (via system instructions) to respond with text that is brief and un-annotated.
|
||||
Reference in New Issue
Block a user