Agentic Support

Edge is built for the agentic era. AI coding agents — like Cursor, Windsurf, Claude and custom toolchains — can deploy, manage and scale infrastructure on the Edge Network autonomously, using a purpose-built Agent API and dedicated access codes.

Agent Access Codes

Agent access codes are separate from regular API keys and are designed specifically for AI agent use. They provide scoped, controlled access to your Edge account.

Each access code can be configured with:

Product scope — restrict access to specific products (Compute, CDN, Storage, DNS)
Permission levels — control what the agent can do: deploy, manage, or delete
Budget limits — set optional monthly spending caps (soft or hard limits)

Access codes use the format ea_live_... and are created from the Edge control panel under Account → Agent Access. The full secret is shown only at creation, so store it securely.

The Agent API

The Agent API is a self-describing API designed for AI agents. It supports composite actions that combine multiple products into a single deployment, reducing the number of steps an agent needs to take.

Composite Deployments

Deploy a static site — creates a storage bucket, uploads files (or clones from GitHub/GitLab), sets up CDN, configures a custom domain with SSL, and adds DNS records — all in one call
Deploy an app — provisions a VM with a setup script, optionally adds CDN and DNS configuration
Update a deployment — uploads changed files with incremental updates, optionally purges CDN cache

Project Management

Every deployment creates a project that groups all related resources. Agents can check project health, scale compute resources, and tear down entire projects when they're no longer needed.

Key Features

Discovery endpoint — agents can query /agent to discover all available endpoints and capabilities
Dry run mode — send X-Dry-Run: true on any mutating request to preview what would happen without making changes
Idempotency — use Idempotency-Key headers for safe retries
Human-readable responses — every response includes a tell_user field with a plain-English summary of what happened

What Agents Can Do

With an agent access code, an AI agent can:

Create, resize and delete virtual machines
Deploy static sites and full-stack applications
Manage storage buckets and upload files
Configure DNS zones and records
Set up CDN deployments with custom domains
Monitor project health and resource usage
Scale infrastructure up or down

Getting Started

Sign in to the Edge control panel
Navigate to Account → Agent Access
Create a new access code with the products and permissions your agent needs
Pass the code to your AI agent as a bearer token

Full documentation is available at edge.network/docs/agent, including guides for deploying static sites, deploying apps, and using the discovery endpoint.