The Agent Stack
Vercel just shipped the full Agent Stack. One framework. Five SDKs. Everything developers need to build production agents without vendor lock-in.
Why it matters
Vercel is positioning itself as the infrastructure layer for agent development, releasing a complete open-source framework (eve) and toolkit that abstracts away model routing, workflow durability, sandboxing, and multi-platform deployment—solving the fragmentation problem that forces teams to either lock into a single vendor or stitch together disparate solutions.
The key facts
9 to knoweve framework launched in public beta (open-source, opinionated agent implementation)
Five core SDKs: AI SDK, AI Gateway, Workflow SDK, Vercel Sandbox, Chat SDK, Vercel Connect
AI Gateway routes across hundreds of models from single endpoint with no vendor markup
Vercel Sandbox provides isolated microVMs with Docker support and credential injection
Vercel Connect in public beta supporting Slack, GitHub, Snowflake, Salesforce, Notion, Linear + OAuth
Chat SDK enables single agent deployment across 12+ channels from one codebase
Workflow SDK provides durability, state persistence, and checkpointing for multi-step agent runs
SERHANT and FLORA cited as production customers running agents on the stack
Vercel leveraging existing infrastructure from billion preview deployments and 6M daily builds
Go to the source
Vercel Blogvercel.com
Publisher excerpt: Agents are designed to do almost any kind of work, from answering support tickets to writing code. No matter how complex the workload, how long it runs, or how many turns it takes to complete, every agent needs three core capabilities to operate: Implementing these capabilities to build a complete…
