Plaid logo
Core Exchange
ALL DOCS

Production access

Open nav
Core Exchange
Close search modal
Plaid.comGet Started

Production access

Request production access, run stealth testing, and launch your integration

Requesting production access

Once all required endpoints have been validated, a production access request button becomes available in the Dashboard.

When a production request is submitted:

  • Plaid performs a final configuration review
  • OAuth and data flows are rechecked
  • Institution branding is confirmed
  • Production timing is coordinated with your team

Your integration is not made available to end users until timing is explicitly confirmed.

For platform integrations, production access is not requested repeatedly. Instead, each new institution added to the platform is submitted as a new integration and approved by Plaid before production traffic is enabled.

Stealth testing (optional)

Stealth testing is a production-only testing phase that runs under a non-discoverable (stealth) institution name.

It allows you to:

  • Test your integration through real third-party applications
  • Send limited traffic through production infrastructure
  • Run friends-and-family or internal testing
  • Identify edge cases not present in pre-production

Once testing is successful, you can update the institution name and branding to make the integration discoverable.

Launching and discoverability

After production access is enabled, your institution becomes available to Plaid-powered applications based on product requirements and search eligibility.

Discoverability depends on enabled products and supported account types.

Pre-launch checklist

Complete before going live:

Configuration

  • OAuth server in production environment
  • Discovery endpoint accessible
  • JWKS endpoint published and accessible
  • Production credentials issued to Plaid
  • mTLS certificates configured (if applicable)
  • Organization name, logo, and URL confirmed

Testing

  • All required endpoints pass validation
  • Error handling verified
  • Token expiration and refresh flow tested

Monitoring

  • Latency monitoring enabled
  • Error rate alerts configured
  • Uptime monitoring in place
  • Log aggregation working
  • On-call rotation established

Documentation

  • Runbooks created for common issues
  • Escalation paths documented
  • Support contact information shared with Plaid
  • Maintenance window procedures defined