`/neuroflow:setup`¶

Interactive credential wizard for MCP integrations.

/setup guides you through connecting the neuroflow MCP integrations — Miro for visual collaboration, Google Workspace, and optional custom LLM providers. It stores credentials securely in .neuroflow/integrations.json, which is git-ignored by default.

When to use it¶

First time setting up credentials (or if you skipped it during /neuroflow)
Updating an existing credential
Checking which integrations are currently configured

What it does¶

Step 1 — Show current status¶

Displays a status table so you know what's already configured:

Integration      Status
─────────────    ──────
PubMed/bioRxiv   ✅ no credentials needed
Miro             ❌ not configured
Context7         ✅ no credentials needed
Custom LLM       ❌ not configured

Step 2 — Miro setup¶

Miro requires a personal access token. The wizard tells you exactly how to get one:

Go to https://miro.com/app/settings/user-profile/apps
Click Create new app (or use an existing one)
Under Token, click Create token — copy the token shown

Paste your Miro access token (or press Enter to skip):
> eyJhbGciOiJSUzI1NiJ9...

Step 3 — Google Workspace CLI setup¶

See the Integrations guide for full details on getting OAuth credentials.

Step 4 — Custom LLM provider (optional)¶

Optionally configure an alternative LLM API endpoint for Claude Code. Skip this step if you use Anthropic directly.

If configuring: - Enter provider name, base URL, API key, preferred model, and proxy port - Credentials are saved to .neuroflow/integrations.json under custom_llm - Non-secret settings (provider, URL, model) can also be synced to your flowie profile for cross-machine use

e-INFRA CZ

Czech academic researchers can use the e-INFRA CZ free LLM API (https://llm.ai.e-infra.cz) via Metacentrum/e-INFRA CZ membership. See the integrations guide for details.

Step 5 — Save credentials¶

Credentials are saved to .neuroflow/integrations.json:

{
  "miro": {
    "MIRO_ACCESS_TOKEN": "eyJhbGciOiJSUzI1NiJ9..."
  },
  "custom_llm": {
    "provider": "einfra",
    "base_url": "https://llm.ai.e-infra.cz/v1",
    "api_key": "<stored locally, gitignored>",
    "model": "qwen3.5-122b"
  }
}

This file is local only

.neuroflow/integrations.json is excluded from git (added to .gitignore by neuroflow). Your credentials are never committed to your repository.

Activating credentials¶

After running /setup, export the env vars in your shell before starting Claude Code:

export MIRO_ACCESS_TOKEN="eyJhbGciOiJSUzI1NiJ9..."

Add to your shell profile (~/.zshrc or ~/.bashrc) so it loads automatically:

echo 'export MIRO_ACCESS_TOKEN="eyJhbGciOiJSUzI1NiJ9..."' >> ~/.zshrc

What is automatic vs manual¶

Step	Automatic	Manual
MCP server processes started	✅ Claude Code launches them via `npx`	—
Miro token entry	✅ Prompted by `/setup`	⚠️ You must create the token in Miro first
Miro OAuth browser login	❌ Not implemented	Use a personal access token instead
Env var export	❌ Not automatic	Run `export …` or add to shell profile

Reminder behavior¶

If you mention Miro and MIRO_ACCESS_TOKEN is missing, Claude offers to run /setup.
You can always re-run /setup to add or update credentials.

Files read and written¶

Direction	Files
Reads	`.neuroflow/integrations.json`
Writes	`.neuroflow/integrations.json`

Integrations guide →
/neuroflow — also offers to run setup on first use

/neuroflow:setup¶