Five AI plugins for Obsidian that keep knowledge from getting lost in your notes

On July 9, 2026, Andrey Menshikov, a senior developer at IBS, published a review on Habr of five AI plugins for the note-taking app Obsidian. In his experience, they help keep knowledge from getting lost in personal notes, articles and meetings.

Why knowledge gets lost without a system

If you don't come back to information, you forget it within a week. The note technically exists, but finding it among chats and files is hard. The author has used Obsidian for more than a year as his working system for notes, ideas, talk preparation and project documentation. He describes five specific AI plugins that have grown into a full ecosystem around Obsidian.

"Technically, the knowledge exists, but where is it?

In the notes? In a chat with someone?" — Andrey Menshikov, senior developer at IBS, in his Habr article.

Which plugins the author covers

• Copilot (github.com/logancyang/obsidian-copilot) — a general-purpose assistant with RAG support that works with OpenAI, Anthropic Claude, Google Gemini, OpenRouter and Ollama; used for finding information, outlining articles, summarizing, and taking meeting notes
• Smart Connections (github.com/brianpetro/obsidian-smart-connections) — semantic search by meaning rather than keywords: it builds embeddings of your notes and finds thematically related entries, includes Smart Chat, and emphasizes local data processing
• Text Generator (github.com/nhaouari/obsidian-textgenerator-plugin) — generates, rewrites, shortens and expands text right in the editor; supports more than 100 providers and models, plus custom prompt templates
• Chat GPT MD (github.com/bramses/chatgpt-md) — saves conversations with the model as markdown files directly in the knowledge base; handy for brainstorming and writing up research
• Local GPT (github.com/pfrankov/obsidian-local-gpt) — runs fully locally through Ollama, so data never leaves the device, which removes NDA and confidentiality concerns for work documentation

How these plugins are used in practice

The author lists six use cases: semantic search instead of keyword search, uncovering hidden connections between notes, quickly processing the incoming flow of information, creating flashcards, regularly reviewing the knowledge base, and talking to your own knowledge base through a RAG chat.

The author concludes that how well all these tools work depends heavily not on the size of the knowledge base but on how atomic and clearly titled the notes are. The plugins handle one giant file worse than a set of small, linked entries.

What this means

The ecosystem of AI plugins around Obsidian has grown to the point where a personal knowledge base stops being just an archive of files. It becomes a system you can hold a conversation with and that finds meaningful connections on its own — provided the user writes atomic notes rather than one big document.

FAQ

Which of these AI plugins for Obsidian are free?

All five — Copilot, Smart Connections, Text Generator, Chat GPT MD and Local GPT — are distributed as open-source projects on GitHub.

Which plugin should you pick if your data can't be sent to external services?

Local GPT: it runs through Ollama entirely on the user's device and doesn't send data to external APIs, which makes it suitable for documents under NDA.

How is

Smart Connections different from regular Obsidian search?

Smart Connections builds embeddings of your notes and searches by meaning rather than exact keyword matches. That lets it find thematically related notes that regular search would miss.