Migration
Migrate from Vercel AI Gateway to Concentrate
Keep one gateway in front of your providers, and access all models through Concentrate with spend views, request logs, fallbacks, redaction, and team controls.
Code change
Base URL + key
Keep your SDK, prompts, and parsing in place.
Rollout
By workload
Move low-risk traffic first, then the rest.
Testing
Compare first
Check output, latency, tokens, and fallback paths before cutover.
New capabilities
What your team gains with Concentrate
01
Keep one gateway
Reach current models across providers through one endpoint.
02
Add management
Get spend by team and key, request logs, API keys, and fallbacks.
03
Keep review ready
Use redaction, SSO, RBAC, and audit logs.
Migration basics
Frequently asked questions
When should a team move from Vercel AI Gateway to Concentrate?
Move when you want spend by team or key, request logs, provider fallbacks, API keys, redaction, and audit logs in one managed request path rather than a routing layer alone.
Can we keep using the Vercel AI SDK?
Yes. You can keep the AI SDK in application code and point it at Concentrate. See the Vercel AI SDK integration for setup and the migration guide for moving the gateway layer.
How should we test the migration?
Replay known requests through both paths and compare outputs, latency, token counts, errors, and fallback behavior. Move traffic by workload instead of all at once.