Using Ollama with Obsidian usually means connecting Obsidian to Ollama’s local API so you can summarize, rewrite, autocomplete, or query your notes without sending them to a cloud service. The common setup is: install Ollama, pull a model, make sure the local server is running on `http://localhost:11434`, then connect it through an Obsidian community plugin such as AI Providers, Local GPT, Copilot, or Companion, depending on the workflow you want.[^1][^2][^3][^4] ## Typical setup 1. Install Ollama and verify it works from the terminal with a model like `ollama run mistral` or another model you prefer.[^3][^1] 2. Confirm the local API server is available at `http://localhost:11434`.[^1] 3. In Obsidian, install a community plugin that supports Ollama, then choose Ollama as the provider and select your model.[^4][^3][^1] 4. Test the connection inside the plugin settings before using it in notes.[^3] ## Good plugin choices - **AI Providers**: good if you want Obsidian to route prompts through Ollama cleanly, with provider-based configuration.[^1] - **Local GPT**: useful for note actions, context menus, and hotkeys.[^1] - **Copilot / autocomplete plugins**: better if you want inline suggestions while typing.[^4][^3] - **Simple Ollama plugins**: some community plugins let you select text and replace it with generated output directly.[^5] ## Practical examples A memoir writer’s use case might be: select a paragraph in a draft note, ask Ollama to tighten the prose, or generate alternate openings while keeping everything local. Another useful workflow is to assign a hotkey for a “rewrite” or “summarize” action so you can process notes quickly while drafting.[^5][^1] ## One thing to watch Some Obsidian plugins require the Ollama server to allow requests from Obsidian’s app origin, especially on macOS, and may need a command like `OLLAMA_ORIGINS=app://obsidian.md* ollama serve` or an equivalent environment setting. If a plugin fails to connect, the usual causes are the wrong API URL, the wrong model name, or Ollama not actually running.[^6][^7][^3] <span style="display:none">[^8][^9]</span> <div align="center">⁂</div> [^1]: https://openresource.dev/articles/running-a-local-ai-inside-obsidian-with-ollama/ [^2]: https://ollama.com/blog/llms-in-obsidian [^3]: https://github.com/j0rd1smit/obsidian-copilot-auto-completion/blob/master/docs/how-to/Ollama setup guide.md [^4]: https://noted.lol/obsidian-companion-ai-plugin/ [^5]: https://www.reddit.com/r/ObsidianMD/comments/1n6d2do/here_is_a_simple_ollama_plugin_for_obsidian/ [^6]: https://www.youtube.com/watch?v=tC9wegnRyZk [^7]: https://www.youtube.com/watch?v=9YYB8a_ehc4 [^8]: https://ai.plainenglish.io/your-obsidian-notes-just-got-smarter-a-personal-journey-with-anythingllm-and-ollama-78cde30d3414 [^9]: https://community.obsidian.md/search?q=ollama