Skip to content
All tool news

Tool desk · Updated daily

Feature·Vercel·

Vercel Makes Eve 1.0 Generally Available for Production Agents.

Vercel makes eve 1.0 generally available as an open-source agent framework with durable workflows, approvals and managed hosting.

CW

Create With tool desk

3 sources checked · 3 min read · 4 sections

ShareLinkedIn
Vercel Makes Eve 1.0 Generally Available for Production Agents

eve 1.0 is now generally available as Vercel’s open-source framework for building production-ready agents. It combines model calls, tool execution, durable workflows, approval controls and isolated code execution in one stack.

According to Vercel’s launch post, the framework has moved from its public preview in June to a stable 1.0 release. Vercel shipped more than 230 releases and merged thousands of pull requests during that period.

The release also makes eve faster, expands its integrations and improves observability. Vercel has worked on making the framework easier for coding agents to build with, as well as for developers building agents directly.

Releases since public preview
230+
shipped between June and general availability
Pull requests merged
Thousands
completed while preparing eve 1.0

What Vercel Added to eve 1.0

At the centre of eve is a harness for calling models and running tools. This handles the repeated loop behind an agent: send context to a model, receive an instruction, run the relevant tool and pass the result back.

Long-running work is handled through the open-source Workflow SDK. It provides durable execution with checkpointing, so an agent run does not have to depend on one uninterrupted process. A workflow can preserve progress between stages rather than starting the whole job again.

The framework also includes built-in multiplayer and agent-to-agent calls. Multiplayer supports workflows involving more than one participant. Agent-to-agent calls allow separate agents to pass work between each other inside a larger process.

Security controls sit around those actions. Builders can choose which actions require approval and control what data gets collected. That creates a defined review point before an agent carries out a sensitive step.

AreaPublic previeweve 1.0
Release statusPublic preview from JuneGenerally available
DevelopmentEarly-adopter testingMore than 230 releases and thousands of merged pull requests
ExecutionFramework under active developmentDurable execution with Workflow SDK checkpointing
ControlsPreview implementationAction approvals and data collection controls
OperationsPreview-stage toolingImproved speed, integrations and observability
Vercel published the eve 1.0 general availability announcement on its product blog.
Vercel published the eve 1.0 general availability announcement on its product blog.

How Builders Can Run eve Agents

An eve project can be hosted on Vercel using Workflows for durable execution and Sandbox for isolated code execution. Sandbox gives generated or agent-selected code a separate environment in which to run. The underlying Sandbox API has also reached a stable release, following the Vercel Sandbox Python SDK 1.0 update.

The pieces cover different parts of an agent run. The eve harness manages model and tool interactions. Workflow SDK coordinates jobs that need to continue across several stages. Sandbox isolates code execution. Approval controls determine when a person must step in.

A builder can put those parts together in this sequence:

  1. 1

    Configure the harness

    connect the model and define the tools available to the agent

  2. 2

    Create the workflow

    split longer work into durable steps with checkpoints

  3. 3

    Set approval rules

    choose which actions must pause for human review

  4. 4

    Isolate code

    send generated or selected code to Vercel Sandbox

  5. 5

    Deploy the agent

    host the workflow on Vercel and inspect it through the available observability tools

Managed hosting is one option rather than a requirement for the framework itself. eve and Workflow SDK are open source, while Vercel provides the associated hosted Workflows and Sandbox services.

From Agent Demo to Managed Workflow

A basic agent can be built as a model call connected to a few tools. Production work adds more moving parts: state must persist, tool failures need to be visible, code may need isolation, and selected actions may need a person’s approval.

eve 1.0 packages those concerns into a single framework. Checkpointing covers durable state. Observability records what happens during a run. Sandbox provides isolated execution. Approval settings put a human decision before selected actions, while agent-to-agent calls support processes split across specialised agents.

The 1.0 label also changes the release status. The framework is no longer presented as a public preview. It is now generally available after the release cycle that began in June.

Frequently asked questions

What is Vercel eve 1.0?

eve 1.0 is an open-source framework for building production-ready agents. It provides a harness for model calls and tool execution, durable workflows, multiplayer support, agent-to-agent calls, security controls and managed hosting options.

When is eve 1.0 available?

eve 1.0 is generally available. It followed a public preview that began in June.

How does eve 1.0 compare with the public preview?

Vercel shipped more than 230 releases and merged thousands of pull requests between the preview and version 1.0. The general release is faster and includes expanded integrations, improved observability and security controls for approvals and data collection.

Does eve 1.0 support human approval steps?

Yes. Builders can choose which agent actions require approval before they run. They can also control what data is collected during agent activity.

Sources

3 checked

How we cover tool news: Create With's tool desk drafts these reports with AI from the sources listed above and checks them against those sources before publishing.

Worth passing on?

ShareLinkedIn

Go deeper on Vercel

Related reading, watching and going.

Everything on Vercel →

The briefing

19 May 2026

How to Migrate Your App from Lovable to Claude Code in 2025

Step-by-step guide to migrating your Lovable app to Claude Code. Learn how to move from Lovable's managed platform to a stack you own and control.

Read →

The briefing

10 Feb 2026

OpenClaw Review: 2 Weeks With This AI Agent

Honest 2-week review of OpenClaw — an autonomous AI agent powered by Claude. We test setup, safety, autonomy, and real value to see if it's worth using.

Read →
34 min

Podcast

My Daughter Built a Website Business, Claude Design First Look & Going All-In on Cloudflare

James and Kieran discuss practical AI-powered development through real-world examples, including a seven-year-old building a website business with Claude Code voice mode, Anthropic's new Claude Design tool, API security challenges in the AI agent era, and Cloudflare's expanding AI infrastructure offerings. The episode covers hands-on experiences with vibe coding, emerging design tools, and the evolving landscape of AI-assisted development.

Watch now ↗

Podcast

53. Why Your Vibe Code Needs Tests, Ditching SaaS & AI Agents That Work While You Sleep

James and Kieran discuss the practical realities of building AI-powered software, emphasizing the importance of testing in vibe-coded applications, exploring alternatives to traditional SaaS tools, and implementing autonomous AI agents for business operations. The episode covers technical topics like Convex databases, app store bottlenecks from AI-generated submissions, and using OpenClaw as a marketing automation agent.

Watch now ↗
38 min

Podcast

52. Your AI Co-Founder: OpenClaw, Manus & the Zero-Person Business

In this Create With podcast episode, James and Kieran are joined by Matt Roberts from Happy Operators for an in-depth discussion on using autonomous AI agents in production. The conversation covers practical experiences with OpenClaw, security considerations, comparison with emerging tools like Manus, and the broader shift toward AI-assisted development. They also discuss the VibeCoding Olympics results, Bubble's new AI capabilities, and research on Claude.MD file effectiveness.

Watch now ↗

Latest tool news

What else changed this week.

All tool news
n8nDigest

What n8n Shipped in Mid-November

n8n shipped a small but notable AI model update for builders working on agents and automation. The latest change adds Gemini 4 Carbon to n8n, giving teams another model option for…

VercelDigest

What Vercel Shipped in Mid-November

Vercel shipped three updates for sandboxed runtimes and AI Gateway users. The changes cover a new Python image, a new model in Token Pass, and a web-search decision model that is…

The Create With Briefing

Don't watch forty changelogs. Read one email.

Every Tuesday: the tool changes worth knowing, real business use cases, and what's on near you. Free, unsubscribe any time.