Connectors & your own keys

Your accounts, wired in

Beyond the built-ins, your app can run on your own services — 33 connectors across payments (Stripe), email (Resend, SendGrid, Mailchimp, Brevo), messaging (Telegram, Twilio, Slack, Discord), data (Airtable, Notion, HubSpot, Google Sheets/Docs/Drive), scheduling (Calendly), scraping (Firecrawl), maps (Google Maps), workflows (n8n), project tools (Linear, GitHub), and AI on your own keys (OpenRouter, OpenAI, Anthropic, ElevenLabs, Deepgram, DeepL, Perplexity, Groq, Grok, Mistral, Replicate, fal.ai).

The flow is one sentence and one paste:

1. Ask for the feature — "take payments with Stripe", "send booking confirmations by email".
2. opix builds the integration with correct, service-specific patterns.
3. You paste the key once under project settings → Connectors. Done.

Browse them all on the integrations page.

How keys are protected

• Keys are stored encrypted (AES-256) in your project's vault.
• They're write-only: after saving, only the last four characters are ever shown again.
• They're injected server-side into your app's environment — never visible in chat, in generated code, or in the browser.
• AI keys used for in-app AI are extra-restricted: they route through the opix gateway and are never placed in your app's environment at all.

Honest degradation

A missing key never fakes success and never blocks a build:

• The app builds and previews normally.
• Features that require the key show a clear, designed setup note, and their API routes answer 503 until the key exists.
• Side-effect sends (notifications, confirmation emails) are skipped while the underlying action still completes.

That means you can build first, decide on services later, and nothing pretends to work before it does.