LLM Hub brings multi-provider chat and automation to Obsidian

LLM Hub is a free, open-source desktop plugin that integrates AI chat, workflow automation and semantic search into Obsidian. It supports Gemini, OpenAI, Anthropic, OpenRouter and Grok alongside local model servers and CLI backends. Users can configure multiple providers and switch models during conversations.

Compatible models can read, create, search and edit vault notes through function calling, with confirmation dialogs for proposed edits and configurable folder restrictions. A visual workflow builder offers 25 node types, while natural-language instructions can generate or modify workflows. Live voice conversations use Gemini or OpenAI independently of the model selected for text chat, with optional read-only note access and RAG.

Semantic search uses a locally stored vector index to retrieve relevant vault content, with Gemini or custom embedding providers. A dedicated search interface supports PDF previews, result filtering and sending selected material into conversations. Separate Dashboard Hub, Discussion Hub and LLM Connect Hub plugins extend these capabilities to dashboards, model discussions, Discord and Kakeratta.

Chat histories and workflow logs can be encrypted independently. Local storage does not mean all processing stays on-device: configured API providers receive messages and attachments, and external search services receive queries. CLI backends lack built-in vault tools and web search, while local tool support depends on the server and model. Workflows should be reviewed because HTTP nodes can transmit vault data externally.