LiquidText
The infinite workspace for deep document analysis and multi-source synthesis.

Real-time RAG and semantic discovery for your local markdown knowledge base.
Smart Connections is the premier Retrieval-Augmented Generation (RAG) implementation for the Obsidian ecosystem. By 2026, it has evolved from a simple link suggester into a sophisticated local vector database management system. It functions by indexing a user's entire markdown vault into a vector space—using either local transformer models or remote embeddings (OpenAI/Cohere)—enabling 'Smart Chat' which interacts with the totality of a user's historical notes. The technical architecture prioritizes privacy-first AI, allowing users to run inference via Ollama or LM Studio locally, or leverage high-end API providers. Its market position is unique as it bridges the gap between static personal notes and dynamic agentic workflows, effectively turning a folder of text files into a structured, queryable data lake. The plugin's 2026 iteration includes advanced features like hybrid search (combining keyword and semantic) and multi-file context injection, making it the industry standard for researchers, developers, and knowledge workers who require deep context awareness without manual internal linking.
Uses Transformers.js to run embedding models directly in the Obsidian environment, ensuring zero data leaves the machine.
The infinite workspace for deep document analysis and multi-source synthesis.
Empower your teams to learn, practice, and perform with AI-driven sales enablement and microlearning.
Transform fragmented datasets into navigable, high-fidelity neural knowledge graphs for RAG orchestration.
The minimalist's gateway to focused reading and intelligent content archival.
Verified feedback from the global deployment network.
Post queries, share implementation strategies, and help other users.
Real-time cosine similarity calculation between the active note and the entire vault index.
Allows injection of global system instructions into the LLM context window.
Combines BM25 keyword matching with vector-based semantic search.
Connects via local API endpoints to models like Llama 3 or Mistral.
Allows users to drag and drop multiple notes into the chat context to synthesize data.
Analyzes note content to suggest and apply YAML frontmatter tags.
Synthesizing themes across hundreds of research PDFs and markdown notes.
Registry Updated:2/7/2026
Finding logic patterns across multiple code-snippet notes without remembering file names.
Identifying recurring mood patterns or habits over years of daily notes.