Obsidianworks with obsidian

Run code and AI inside your Obsidian vault.

Znote opens your vault as-is — wikilinks, embeds, attachments, your settings. It adds executable code blocks and inline AI, and saves the results as plain Markdown you can read right back in Obsidian.

Download free →
✓ Obsidian plugin included ✓ No vault changes ✓ Plain .md, always
A note in Obsidian with an executable barChart code block — the Znote companion plugin adds copy and ▶ Run in Znote buttons
Your vault in Obsidian — every executable code block gets a "▶ Run in Znote" button.

Your notes learn to run.

Write a JS, Node or shell block in any note of your vault, press play — the chart or table renders inline. The result is saved as plain Markdown, so Obsidian can read it back.

a
Run code in any note
Query a database, call an API, parse a file — press play. Save the note: it's a tool now.
b
Results readable in Obsidian
Text output is saved as a plain ```output block; charts as a PNG in your attachments folder. With the plugin, they render as RESULT panels in reading view.
c
A one-click companion plugin
Znote proposes it the first time you open your vault — one click, done. It adds "Open in Znote" and a hover "▶ Run in Znote" button inside Obsidian, plus znote://open deep links.
api-status.md Znote
▶ Run
const r = await fetch(api + '/health')
console.log(r.status, await r.text())
```output
200 OK · all services up
```
result written below the fence — plain Markdown, readable anywhere
api-status.md Obsidian · reading view
▶ Run in Znote
const r = await fetch(api + '/health')
console.log(r.status, await r.text())
result
200 OK · all services up
the same ```output block — rendered as a RESULT panel by the plugin
Same note — the raw Markdown Znote saves (left), Obsidian's reading view (right).

Your vault stays exactly as it is.

Your notes stay plain .md files — Znote never rewrites what you didn't change.
Your Obsidian settings are respected: inline title, line width, attachment folder.
Wikilinks, embeds and images work the Obsidian way.
No extra files added to your vault.
Stop using Znote tomorrow — your vault won't notice.

AI that answers from your vault.

Type @ to hand the AI a note, a template, a code result. Type / for prompts — the answer lands inline, in the note you're writing.

a
Your notes as context
@user-story-template → "turn this brainstorm into a user story."
b
Every prompt is a note in your vault
/brainstorm, /rewrite, /specify — each one a Markdown file you can override, or add your own.
c
Private by design
Bring your own key, or run 100% local with Ollama.
brainstorm.md Znote ran 0.9s
Brainstorm — gym finder app

I want make app for people find gym close. Right now hard to know where gyms are. Also booking for class complicated.

user-interviews.md
research
user-story-template.md
templates
competitor-research.md
research
Turn this idea into a user story, following @user-story-template.
User story — find a gym nearby
As a gym-goer, I want to see gyms open near me on a map, so that I can pick one without calling around.
✓ Map shows open gyms within 2 km
✓ Each gym links to one-tap class booking
Your files in, answer inline — no chat window, no copy-paste.

Open your vault. Nothing to migrate.

Free, local, plain .md. Point Znote at your vault folder — that's the whole setup.

Frequently asked.

No. Notes stay plain Markdown, settings stay yours. The only things Znote writes are the ones you ask for: code output blocks and result snapshots — and re-running or clearing a block cleans those up.
Yes, side by side on the same vault. Each app reloads changes made by the other.
No — output is a standard fenced code block, and snapshots use Obsidian's own embed syntax. With the plugin they render as RESULT panels; without it they're still clean, readable Markdown.
No. Obsidian is where your knowledge lives; Znote makes it work — code execution and inline AI, results saved back into your notes. Keep both open.
Yes — the app and every feature are free, with free AI requests included. A one-time 29.90€ license unlocks unlimited AI with your own models.
Still have questions? See full FAQ →