Integration
Use the Vercel AI SDK with Concentrate
Send Vercel AI SDK traffic through Concentrate to access models, track spend, and keep provider routes manageable.
App
Next.js
AI SDK calls routed through Concentrate.
Model
GPT-5.5
Current model access through one API.
Review
Logs
Debug request status and spend.
New capabilities
What your team gains with Concentrate
01
Keep SDK ergonomics
Use familiar Vercel AI SDK patterns while routing through Concentrate.
02
Track production usage
See spend, request logs, models, providers, and teams.
03
Use current models
Access current GPT, Claude, Gemini, DeepSeek, Qwen, and other models.
Integration basics
Frequently asked questions
Does Concentrate replace the Vercel AI SDK?
No. Keep the Vercel AI SDK for app-side generation, streaming, UI, and tool-call patterns. Concentrate sits behind the model call so provider routes, keys, spend, logs, and fallbacks do not have to live in each Next.js route.
What changes when a Vercel AI SDK app uses Concentrate?
Usually the API key, base URL, and model route. The app can keep its SDK structure while Concentrate records which team, key, model, provider, token count, status, and spend were used.
Should we test streaming and tool calls before switching traffic?
Yes. Move one low-risk endpoint first, then compare streaming shape, tool-call payloads, JSON output, error handling, latency, and token counts before moving customer-facing chat or agent flows.