Get started

Up and running
in 60 seconds.

Choose how you use Claude. The fastest path is Claude Desktop — one URL, no config files.

Claude Desktop
Claude Code
Manual (CLAUDE.md)
1

Get your API key

You should have received your key by email. It looks like this:

e5a93009-8ad9-4b44-9f6f-840d9c8c32da

No key yet? Request access →

2

Add to Claude Desktop

Open Claude Desktop → Settings → Integrations → Add Integration

Enter the styr-ai URL:

https://app.savageroar.se/api/mcp-server

When prompted for authentication, enter your key as a Bearer token.

ℹ️ Claude Desktop will automatically discover all memory tools and make them available in every session.
3

Verify it works

Start a new session and type:

Call read_memory and show me my project state.
✓ If Claude returns your project memory, you’re live. All tools are active.
1

Add via CLI

Run this once in your terminal:

claude mcp add styr-ai \ --transport http \ --url https://app.savageroar.se/api/mcp-server \ --header "Authorization: Bearer YOUR_KEY"
2

Verify

Start a Claude Code session and run:

claude mcp list

You should see styr-ai listed as active.

✓ All memory tools are now available in every Claude Code session.
1

Add to CLAUDE.md

Create or open CLAUDE.md in your project root. Add:

## Persistent Memory Memory API: https://app.savageroar.se/api/mcp Key: YOUR_API_KEY ## Session Start Call read_memory. Present the last session summary, active decisions, and recent learnings. ## Session End Call write_session with: summary, changes[], next_steps[], project_phase, energy.
2

Verify

Start a new session and say: “Call read_memory and tell me where we left off.”

✓ Claude will call the API and return your project state.

Test your connection

Paste your API key to verify it's active:

All set?  Open your memory dashboard →