Obsidian MCP Server

# Future Improvements Ideas for future development of obsidian-mcp-server. --- ## Migration Tool **Status**: Deferred **Description**: Tool that automatically generates `.agent/vault.yaml` from an existing vault structure. **Implementation ideas**: - Scan root folders and detect naming patterns - Identify templates folder (look for `.md` files with Mustache/placeholder syntax) - Detect tag registry files - Present detected structure to user for confirmation - Generate complete `vault.yaml` **Tool signature**: ```python @mcp.tool() def generar_vault_config(confirmar: bool = False) -> str: """ Escanea el vault y genera .agent/vault.yaml. Args: confirmar: Si True, crea el archivo. Si False, solo muestra preview. """ ``` --- ## Multi-Vault Support Support connecting to multiple vaults simultaneously. --- ## Vault Templates Repository Pre-made vault configurations for common use cases: - Personal Knowledge Management - Academic Research - Software Documentation - Creative Writing --- ## Automated Image Captioning (VLM) **Status**: Proposed **Description**: Use local Multimodal models (Ollama + Llava/Moondream) to automatically generate descriptions for images that lack them. **Implementation ideas**: - New tool `auto_caption_images(path, use_model=True)`. - Scan vault for images without matching wikilinks captions. - Send image to Ollama API for description. - Either edit the note to add the caption `![[img|Generated Description]]` or inject purely into metadata. - Benefits: Full semantic search for images without manual effort.