Repolode
Submit
TarsLab

TarsLab/obsidian-tars

An Obsidian plugin that generates text in your notes from tag-triggered AI assistants across many LLM providers, desktop and mobile.

Open-source alternative to:Notion AI

View on GitHub
Stars
336
Forks
30
Open issues
2
Last commit
2 months ago

What is it?

Tars is a plugin for Obsidian, the Markdown note-taking app. You write a message in a note, tag it with a user or assistant tag such as #User or #Claude, and the plugin sends the conversation to the configured model and writes the reply back into the note. Conversations are structured with blank-line-separated paragraphs, so a note can hold a multi-turn exchange with a model.

The plugin connects to a wide range of providers, including OpenAI, Claude, Gemini, DeepSeek, SiliconFlow, OpenRouter, Ollama, Kimi, Doubao, Qwen, Zhipu and Qianfan, and works on both desktop and mobile. It also handles image input for embedded images and PDF analysis on supported models, and can export conversations as JSONL datasets for fine-tuning with ms-swift.

Key features

  • 01Trigger AI replies by tagging a paragraph with an assistant tag such as #Claude.
  • 02Works with many providers, including OpenAI, Claude, Gemini, DeepSeek, Ollama and Chinese model services.
  • 03Supports image generation and editing with GPT-Image-1.
  • 04Analyzes embedded images and PDFs on models that support vision or documents.
  • 05Exports conversations to JSONL datasets for ms-swift fine-tuning.
  • 06Provides tag autocomplete and custom prompt templates.

Use cases

  • Chatting with a model inside your notes while keeping the conversation saved in the vault.
  • Comparing answers from different providers by tagging the same prompt with different assistants.
  • Building fine-tuning datasets from conversations you have already recorded in Obsidian.
  • Using a local Ollama model for private note-based writing and Q&A.

Quick start

  1. 1

    Install dependencies

    npm install
  2. 2

    Start the watch build for development

    npm run dev
  3. 3

    Build the production bundle

    npm run build

Things to know

  • Requires an API key or a running model provider (such as Ollama) before the assistants will respond.
  • Image and PDF analysis only works with embedded files, not external URLs.
  • Model lists come from each provider's API, so some models may not appear in the picker.

In collections

Claude Code tools, skills & pluginsRun LLMs locallyStatic site generators & documentation tools

Topics

#obsidian-plugin#llm#chatgpt-client#markdown#note-taking#ollama#claude#openai#deepseek#doubao

FAQ

What is obsidian-tars?
An Obsidian plugin that generates text in your notes from tag-triggered AI assistants across many LLM providers, desktop and mobile. Tars is a plugin for Obsidian, the Markdown note-taking app. You write a message in a note, tag it with a user or assistant tag such as #User or #Claude, and the plugin sends the conversation to the configured model and writes the reply back into the note. Conversations are structured with blank-line-separated paragraphs, so a note can hold a multi-turn exchange with a model.
Is obsidian-tars free to use? What license does it have?
obsidian-tars is open source under the MIT license. Check the license terms before using it in commercial products.
Is obsidian-tars still maintained?
Updated within 6 months. The last commit to obsidian-tars was on Aug 28, 2026.
What can I use obsidian-tars for?
Chatting with a model inside your notes while keeping the conversation saved in the vault. Comparing answers from different providers by tagging the same prompt with different assistants. Building fine-tuning datasets from conversations you have already recorded in Obsidian. Using a local Ollama model for private note-based writing and Q&A.
How do I install obsidian-tars?
Install dependencies: npm install Start the watch build for development: npm run dev Build the production bundle: npm run build
Where can I get obsidian-tars?
The source code and installation instructions are on GitHub: https://github.com/TarsLab/obsidian-tars

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects