Model Configuration

WiseMindAI's AI features require model configuration first. There are two ways:

1. During the first launch of WiseMindAI, you'll be guided to install local models

2. You can also connect API keys from third-party model providers you already use, such as ChatGPT, DeepSeek, Gemini, and others

Open "Settings" from the lower-left corner of WiseMindAI, then go to "Model Settings" to configure models for different scenarios.

All subsequent model settings can be modified here.

Configure as Needed

1. Local models: Manage local model services such as Ollama and LM Studio. Recommended for offline use or privacy-sensitive materials.
2. General model: Used for AI chat, document summaries, knowledge base summaries, knowledge cards, mind maps, plugins, and other core features. See LLM Key.
3. Embedding model: Converts documents, web pages, notes, and other content into searchable knowledge indexes. This is important for knowledge base Q&A. See Embedding Key.
4. OCR model: Extracts text from images, screenshots, and scanned materials so they can be summarized and used in Q&A. See OCR Text Recognition Service Key.
5. Audio and video transcription model: Extracts text from audio and video materials for summaries, outlines, and knowledge card generation. See Transcription Service Key.
6. Image generation model: Generates infographics, posters, and other images from knowledge points or prompts. See Image Generation Model Key.

Suggestions

If you are just getting started with WiseMindAI, keep the default local model first, then add one general model key if needed. This makes core features such as document summaries, knowledge base Q&A, and knowledge cards more stable.

If you already have keys for OpenAI, Gemini, DeepSeek, Zhipu AI, Alibaba Cloud Bailian, OpenRouter, or other platforms, enter them under the corresponding provider. If you use an OpenAI-compatible endpoint, choose the custom OpenAI-compatible service.