Setting Up AI for mOperator

mOperator needs an AI model to work. You have two options for connecting one:

AI Gateway (Recommended)	Direct API Key
Setup	One Vercel API key	One provider-specific key
Switch models	Change an env var	Change code + keys
Providers	Anthropic, OpenAI, and more	One at a time
Logging	Built-in usage dashboard	Roll your own
Best for	Teams, production, flexibility	Quick start, single model

Pick whichever fits your situation. You can always switch later.

---

Option A: Vercel AI Gateway (Recommended)

AI Gateway is a proxy that routes your AI requests to any supported model provider. One key gives you access to Claude, GPT-4o, and future models. When a better model comes out, you change one env var — no refactoring, no new API keys, no code changes.

Why AI Gateway?

• Model flexibility — Switch from Claude to GPT-4o (or the next big model) by changing AI_PROVIDER in your env vars. No code changes.
• One key to manage — Instead of juggling API keys for Anthropic, OpenAI, and whatever comes next, you manage one Vercel token.
• Built-in logging — See usage, costs, and request history in your Vercel dashboard.
• Future-proof — As new AI providers and models launch, AI Gateway adds support. Your app just works.

Step 1: You Need a Vercel Account

If you followed the Deploy to Vercel guide, you already have one. If not:

1. Go to vercel.com
2. Click "Sign Up" (free tier works)

Step 2: Create an API Key

1. Go to your Vercel dashboard at vercel.com/dashboard
2. Click your avatar in the top-right → "Settings"
3. Navigate to "Tokens"
4. Click "Create Token"
5. Name it mOperator, set scope to your account
6. Copy the token — you won't see it again

Step 3: Add to Environment

Add to your .env.local:

AI_GATEWAY_API_KEY=your-vercel-token-here
AI_PROVIDER=anthropic

That's it. The default model is claude-sonnet-4-5-20250929. To use OpenAI instead:

AI_GATEWAY_API_KEY=your-vercel-token-here
AI_PROVIDER=openai

Optional overrides:

# AI_GATEWAY_URL=https://ai-gateway.vercel.sh   # Default, rarely needs changing
# AI_MODEL=claude-sonnet-4-5-20250929            # Override specific model

---

Option B: Direct API Key

If you just want to get running with a single provider and don't need the flexibility of AI Gateway, you can plug in an API key directly.

For Anthropic (Claude)

1. Go to console.anthropic.com
2. Sign up or log in
3. Go to "API Keys" and create a new key
4. Add to .env.local:

ANTHROPIC_API_KEY=sk-ant-xxxxxxxxxxxxx
AI_PROVIDER=anthropic

For OpenAI (GPT-4o)

1. Go to platform.openai.com
2. Sign up or log in
3. Go to "API Keys" and create a new key
4. Add to .env.local:

OPENAI_API_KEY=sk-xxxxxxxxxxxxx
AI_PROVIDER=openai

Note

With direct keys, switching providers means getting a new API key from a different provider and updating your code/config. AI Gateway removes this friction — but either approach works.

---

Test the Connection

Start your app:

npm run dev

Test via CLI:

npm run cli

Type something like:

hello, are you working?

If you get a response, you're connected.

Choosing a Model

mOperator defaults to these models:

Provider	Default Model	Good For
Anthropic	claude-sonnet-4-5-20250929	Tool use, structured data, long context
OpenAI	gpt-4o	General purpose, fast responses

To override the model, set AI_MODEL in your .env.local:

AI_MODEL=claude-sonnet-4-5-20250929

Troubleshooting

"No AI key configured"

- Make sure you've set either `AI_GATEWAY_API_KEY` or `ANTHROPIC_API_KEY` / `OPENAI_API_KEY` - Restart the dev server after changing env vars

"401 Unauthorized"

- Verify your API key is correct (no extra spaces or line breaks) - For AI Gateway: check the token hasn't been revoked in Vercel settings - For direct keys: check the key is active in the provider's console

"Model not found"

- Check `AI_PROVIDER` is set to `anthropic` or `openai` - If using a custom `AI_MODEL`, verify the model name matches what the provider supports

Cost

• Vercel AI Gateway itself is free — you pay for model usage
• Anthropic Claude: ~$3 per million input tokens, ~$15 per million output tokens
• OpenAI GPT-4o: ~$2.50 per million input tokens, ~$10 per million output tokens
• Typical cost for a small team: $5–20/month depending on usage

Next Steps

1. Create a Slack App so your team can talk to mOperator
2. Optionally connect Salesforce, project management, or GitHub