AI Note-Taking Plugins for Obsidian and Logseq, Compared

Obsidian and Logseq are both note-taking apps built around local markdown files, and both have community plugin ecosystems that add AI features on top of that local-first foundation. Obsidian plugins install from inside the app itself, through its community plugin browser or a beta-installer like BRAT. Logseq has its own plugin system too, but the tool covered in this guide — logseq-copilot — is not a Logseq plugin at all; it is a browser extension you install separately in Chrome or Firefox that then connects to a running Logseq instance.

Match the plugin to the specific problem you want solved, not to a general idea of "AI notes." Each of the five tools below is built around one core job.

Smart Connections is an Obsidian plugin that finds notes semantically related to what you are writing or reading, going beyond Obsidian's built-in exact-text backlinks. It builds embeddings of your notes and shows a "related notes" view alongside your editor, plus a separate chat pane you can use to ask questions with your vault as context. It supports both a bundled local embedding model that runs with no external setup, and connections to external API providers for embeddings and chat if you prefer those instead.

Copilot for Obsidian is a chat plugin that adds a conversational sidebar supporting multiple AI providers, plus a Vault QA mode that retrieves relevant passages from across your entire vault to answer a question. It supports connecting to providers including OpenAI, Anthropic, Google, and Azure, as well as local models through Ollama or LM Studio, so you are not locked into a single vendor. Beyond open-ended chat, it also supports custom prompt templates and commands you can trigger on selected text or the active note.

Text Generator is an Obsidian plugin built to generate AI text and insert it directly into a note, driven by customizable templates and prompts, rather than through an open-ended chat conversation. It supports OpenAI and local models through Ollama, LM Studio, and LocalAI. The core workflow is different from a chat plugin: you select a template (or write your own prompt), run it against the open note or selected text, and the output is written into the note itself.

logseq-copilot is a browser extension for Chrome and Firefox, not an Obsidian plugin and not a Logseq plugin installed through Logseq itself. It runs inside your browser and connects to a Logseq graph running locally, adding an AI chat side panel and the ability to summarize the web page you are reading directly into your Logseq notes. Because it is a browser extension, it is the odd one out in this comparison: it only makes sense if you use Logseq, and installing it will do nothing useful for an Obsidian vault.

BMO Chatbot is a lightweight Obsidian plugin that adds a chat sidebar for talking with a local or cloud model, with the open note available as reference context. It supports local models via Ollama and LM Studio as well as several cloud providers, and it emphasizes a smaller, privacy-first setup over a large feature surface. It does not build a vault-wide embedding index and does not offer template-driven generation — its scope is deliberately narrower than Copilot for Obsidian or Smart Connections.

The table below lines up all five tools on app, install method, core job, and local-model support — use it to confirm your pick before installing.

The right combination depends on the size of your vault, how much you write versus search, and whether you use Obsidian or Logseq at all.

Most disappointment with these tools traces back to picking one for a job it was not built to do, rather than an actual defect in the plugin.

Each tool has real limits worth naming directly rather than glossing over, whichever one you choose.