Skip to content

The AI-native knowledge base for Windsurf

Cascade is fast inside the codebase. Hjarni is where the reasoning lives. Connect the two and your editor stops guessing.

Free to start. No credit card required.

What this unlocks

Workflows that actually use your context

Plan from real constraints, not defaults

Ask Cascade to plan a billing refactor. Before it writes the plan, it pulls your Hjarni notes on the current payment provider, the past migration attempts, and the SLA. The plan reflects your reality.

Generate code that fits the codebase

Cascade reads your Hjarni conventions notes before writing. Result: code that matches how the rest of your repo handles errors, tests, and naming. Less review back-and-forth.

Capture the migration lesson while it's fresh

After Cascade finishes a tricky migration, tell it to write a Hjarni note: what worked, what broke, what to do differently next time. The next migration starts from that note.

Audit-trail every Cascade decision

Have Cascade log architectural decisions to a Hjarni 'decisions' folder during a session. Six months later you can read why the choice was made, in plain Markdown.

One knowledge base across editors

Move between Windsurf, Cursor, and Claude Code on the same project. They all read the same Hjarni notes, so the assistant in each editor speaks the same language about your code.

Setup

Connect Windsurf in about two minutes

  1. 1

    Sign up for Hjarni and add a few notes describing how the codebase is meant to fit together.

  2. 2

    Open Windsurf settings. Find the Cascade section and the MCP servers panel.

  3. 3

    Click Add server. Use the JSON above, or fill the form: name hjarni, URL https://hjarni.com/mcp, transport HTTP.

  4. 4

    Restart Cascade. Windsurf opens the Hjarni OAuth page in your browser.

  5. 5

    Open Cascade and ask it to list your Hjarni notes. Tools appear under the server name you chose.

Windsurf MCP servers config

{
  "mcpServers": {
    "hjarni": {
      "serverUrl": "https://hjarni.com/mcp"
    }
  }
}

The Hjarni MCP endpoint is https://hjarni.com/mcp. The connection uses OAuth, so you sign in with your Hjarni account when prompted. No API key files to manage.

Why Hjarni outside Cascade's workspace

Cascade reads files in your workspace when you open them. That works for code. It does not work for design docs in someone's Notion, decisions in a Linear comment, or the runbook a teammate keeps in their Apple Notes. Hjarni is one place all of that can land, with MCP search built in.

Windsurf's planning step is the highest-leverage moment to inject context. If the plan is wrong, the execution is wrong. Folder-level AI instructions in Hjarni let you steer Cascade's plans per project: 'in the Payments folder, never propose schema changes without an ADR'.

Cascade is one editor. Engineers use more than one. Hjarni stays put while editors change underneath it, so the knowledge base you build today is still useful when Windsurf updates, when you try a new editor, or when a new teammate joins.

Common questions

Questions before you connect Windsurf

Where do I add Hjarni in Windsurf?

Open Windsurf settings, find the MCP servers panel, and add Hjarni. Or edit the mcp_config.json file Windsurf writes for you and add the Hjarni entry directly.

Does Windsurf's Cascade agent use MCP tools?

Yes. Cascade treats MCP tools the same way it treats built-in tools. Once Hjarni is connected, Cascade can search your notes, create new ones, and link them in the same flow.

What should I keep in Hjarni for Windsurf?

The context that is not in your code: architecture rationales, decisions you reversed, runbooks, environment quirks, and team conventions. Cascade reads those when it needs them.

Will Hjarni slow Windsurf down?

No. MCP search runs only when Cascade asks for it, and Hjarni responds in well under a second for typical queries. There is no background sync of your repo.

Can I share a Hjarni knowledge base with my team in Windsurf?

Yes. Hjarni's team plan gives each engineer their own Windsurf connection to the same shared folders. Every Cascade session reads from the same source of truth.

Give Windsurf a memory

Write notes once. Windsurf reads them across every conversation.

Write once. You both remember.

Free to start. No credit card required.

Give your AI a memory

Works with Claude and ChatGPT today. Gemini coming soon.