NOTEGEN.Model configuration
Add OpenAI-compatible providers and manage multiple models under each provider.
Configuration structure
A provider stores its base URL, API key, proxy, and headers, and can contain multiple models. Start from a ChatGPT, Gemini, Ollama, or LM Studio template, or create a custom provider.
Model Platforms
Manage AI platforms, credentials, and models for chat, organization, knowledge, and audio.
Provider fields
| Field | Description |
|---|---|
| Name | Identifies the provider; unique names are recommended |
| Base URL | OpenAI-compatible URL through the version segment, such as https://api.openai.com/v1 |
| API Key | Provider credential; use any non-empty value when a local service does not validate it |
| Proxy | Inherit the global proxy, connect directly, or use a provider-specific proxy |
| Custom headers | Extra HTTP headers required by the provider |
Do not append /chat/completions or /embeddings; NoteGen selects the endpoint for each model type.
Model list
Add one or more models to a provider. Supported types are:
chatfor conversation, Agent, writing, and vision-capable chat;imageandvideofor corresponding generation APIs;ttsfor text to speech;sttfor speech to text;embeddingfor vectorization;rerankfor knowledge-base result reranking.
Fetch model names from the provider when supported, or enter one manually. The type must match the API capability.
Advanced parameters
- Temperature / Top P: normally adjust one, not both.
- Streaming: display output as it arrives when supported.
- Maximum output tokens: leave empty for the provider default.
- Token limit parameter: use
max_completion_tokensby default ormax_tokensfor older compatible APIs. - Voice: TTS models can specify a provider voice such as
alloy.
Run Check Connection after saving. A successful provider check does not guarantee every model type works, so test Chat, embeddings, and speech in their actual features.