--- title: Agent Frameworks on Vercel product: vercel url: /docs/frameworks/agents canonical_url: "https://vercel.com/docs/frameworks/agents" last_updated: 2018-10-20 type: conceptual prerequisites: - /docs/frameworks related: - /docs/ai-gateway - /docs/sandbox - /docs/frameworks/agents/eve - /docs/frameworks/agents/mastra - /docs/frameworks/agents/flue summary: Build AI agents with eve, Mastra, Flue, and Vercel. Learn how each framework connects to Vercel compute, models, and storage. install_vercel_plugin: npx plugins add vercel/vercel-plugin --- # Agent Frameworks on Vercel Agent frameworks coordinate model calls, tools, and application state. You can use them with Vercel to serve agent requests, access models through [AI Gateway](/docs/ai-gateway), and run isolated code with [Vercel Sandbox](/docs/sandbox). ## Related pages > **For AI agents:** Follow these links to understand how this page connects to the rest of the Vercel ecosystem. For the full cross-link map (inbound, outbound, prerequisites, and semantic neighbors), see the .graph.md link below. - [The Agent Stack](https://vercel.com/blog/agent-stack) - [Building an agent with OpenAI Agents SDK and Vercel Sandbox](https://vercel.com/kb/guide/building-an-agent-with-openai-agents-sdk-and-vercel-sandbox) — Learn how to build an agent with with OpenAI Agents SDK and Vercel Sandbox - [Give your eve agent secure access to your private AWS RDS database](https://vercel.com/kb/guide/give-eve-agent-secure-access-to-aws-rds-database) — Connect an eve agent to a private AWS RDS database using Vercel Secure Compute and VPC peering, with a read-only query t - [Deploy to Vercel](https://eve.dev/docs/guides/deployment/vercel) — Deploy an eve agent with Vercel Workflow, Sandbox, Cron, and project credentials. - [Using coding agents to procure Vercel Marketplace integrations](https://vercel.com/kb/guide/using-coding-agents-to-procure-vercel-marketplace-integrations) — Coding agents can now discover, provision, and manage third-party services from the Vercel Marketplace using the Vercel - [How to build a GitHub agent with eve and GitHub Tools](https://vercel.com/kb/guide/github-agent-eve) — Build a GitHub agent with eve, GitHub Tools, and Vercel Connect. Register AI-callable GitHub tools, gate writes behind d - [Running eve Agents Durably on Vercel](https://vercel.com/docs/agent-harnesses/eve) — eve agents are durable out of the box on Vercel, with no extra configuration. - [Agent Resources](https://vercel.com/docs/agent-resources) — Set up AI coding tools with Vercel documentation, reusable skills, and secure access to projects, deployments, and logs. Full cross-link map for this page: [/docs/frameworks/agents.graph.md](/docs/frameworks/agents.graph.md) Choose a framework guide to set up a project and run it locally. Each guide explains its Vercel integrations and hosting requirements, including deployment steps where the framework supports running on Vercel. | Framework | Guide | | --- | --- | | [eve](/docs/frameworks/agents/eve) | Build a durable agent with Vercel-native deployment and session state. | | [Mastra](/docs/frameworks/agents/mastra) | Deploy an agent API with Mastra's Vercel deployer and connect it to AI Gateway. | | [Flue](/docs/frameworks/agents/flue) | Run an agent with Flue and connect its model calls and code execution to Vercel. | If you're choosing between eve and building an agent directly with AI SDK, see [Choosing between eve, AI SDK agents, and `WorkflowAgent`](/kb/guide/eve-vs-ai-sdk-agents-vs-workflowagent). Compare execution state, approvals, and channels to decide how much your application needs to manage. ## Vercel services for agents Your framework determines how an agent runs and manages its state. Vercel provides the infrastructure you can connect to it: - [Vercel Functions](/docs/functions) handle incoming requests and scale with traffic. Agent work running inside a function must fit within its [duration limit](/docs/functions/limitations#max-duration). - [AI Gateway](/docs/ai-gateway) routes model requests through a single endpoint with usage tracking and provider fallbacks. - [Vercel Sandbox](/docs/sandbox) isolates generated code and command execution from your application. - [Vercel Daemons](/docs/daemons) run background workers that continuously pull jobs from a queue or workflow engine. See the framework guide for worker requirements. - [Vercel Passport](/docs/passport) protects internal web applications with your identity provider. Use verified visitor identity to authorize agent requests and access to conversation history. ## Persist execution, workspaces, and files Choose persistence based on what your agent needs to retain: | Need | Vercel service | Example | | --- | --- | --- | | Resume multi-step work after interruptions | [Vercel Workflows](/docs/workflows) | Run an agent, wait for approval, then publish its result. eve uses Workflows for session execution on Vercel. | | Reuse a working directory across sandbox runs | [Vercel Drives](/docs/sandbox/concepts/drives) | Keep source files, datasets, or a dependency cache and mount them in a later sandbox. | | Store and serve files outside the sandbox | [Vercel Blob](/docs/vercel-blob) | Save uploaded documents or generated reports. eve's file memory provider also uses a private Blob store. | These services can work together: a workflow coordinates an agent that edits files on a Drive and saves its finished report to Blob. Drives mount inside Vercel Sandbox; Blob exposes files through its SDK and HTTP delivery. Framework-specific conversation history and execution state can require a separate database. Follow each framework guide's storage requirements and use [Marketplace storage integrations](/marketplace/category/storage) when you need a remote database. For agents that connect to external services or chat platforms, see [Vercel Connect](/connect) and [Chat SDK](/docs/chat-sdk). --- [View full sitemap](/docs/sitemap) If you encounter incorrect information, can’t find what you need, need a capability Vercel doesn’t currently offer, or have any other feedback to report, please do it by following the instructions at https://vercel.com/feedback.md.