Question 1

Can I run it on Ollama or the OpenCode local server and still have it edit my notes?

Accepted Answer

No. Ollama, the OpenCode local server, and the CLI backends (Antigravity, Claude Code, Codex) use marker mode rather than function calling, so chat cannot read or edit notes directly. Gemma 4 also drops vault tools whenever RAG or web search is on. LM Studio, vLLM, and AnythingLLM do get vault tools if the model handles OpenAI-style function calling, and incompatible local models are auto-detected on first use and downgraded.

Question 2

How do I actually enable web search?

Accepted Answer

Pick a supported official-host API model, then check Web search in the search menu beside the model picker; prompt wording alone does not enable it. One semantic-search setting can run alongside it so vault context joins web results.

Question 3

Can I use it without buying API keys?

Accepted Answer

Yes. CLI options use account authentication instead of keys, Gemini offers per-model daily-limit switching, and local LLMs plus RAG run without a hosted key. Multiple providers can stay configured at once.

Question 4

When do I need workflows instead of chat?

Accepted Answer

Chat handles interactive Q&A against the vault, while the Workflow Builder automates multi-step tasks with a visual editor and 25 node types, encrypted execution logs, and export paths into Dashboard Hub and notes.

Question 5

What are the main features?

Accepted Answer

Multi-provider LLM chat, vault operations that read, write, search and edit notes, a workflow builder with 25 node types, local semantic search (RAG) with a PDF preview, image generation, edit history with a diff view, and optional password protection for chat history and workflow execution logs.

Question 6

How many providers can I configure at once?

Accepted Answer

Several simultaneously, each with its own API key and settings. Supported providers include Gemini, OpenAI, Anthropic, OpenRouter, Grok, OpenCode Zen and Go, local LLMs such as Ollama, LM Studio, vLLM and AnythingLLM, and CLI backends like Antigravity, Claude and Codex. You can switch models freely during chat.

Question 7

Can it edit my notes directly?

Accepted Answer

Vault operations use function calling for Gemini, OpenAI, Anthropic, OpenCode Zen and Go, and for tools-capable local LLMs through LM Studio, vLLM or AnythingLLM. Ollama and the OpenCode local server run in marker mode instead, and CLI backends do not support vault tools, so those use Workflows or the content and selection variables.

Obsidian LLM Hub

About this project

Highlights

Questions about this entry

More Jev resources like this

From the Jev community