Skip to content

Developers

Build on Edge from wherever you write code

Human documentation, machine-readable specifications and agent tooling, all published at predictable URLs. Use the CLI, the REST API, an S3 SDK or an AI agent: it's the same platform underneath.
Create free accountRead the docs

No card required. Agents can even sign themselves up.

Start here

Four ways in

Readable documentation for every surface: the products, the REST API, the Agent API and the CLI.

Start in one line

Install, discover or download

Install the CLI, point an agent at the self-describing API, or pull the spec into your own tooling.

Prefer npm? npm install -g @edge-network/cli

terminal
# Install the Edge Network CLI
curl -fsSL https://edge.network/install.sh | sh

# Or point any agent at the self-describing API
curl https://edge.network/agent

# Or fetch the OpenAPI spec
curl https://edge.network/openapi.json

Edge CLI

Every product from your terminal

A single binary with no dependencies. Manage VMs, CDN, DNS, storage and Shield, with table, JSON or YAML output for scripting.
  • macOS, Linux and Windows, amd64 and arm64
  • Set EDGE_API_KEY in CI and pipelines
  • Shell completion for Bash, Zsh and Fish
  • CLI actions tagged in your activity log
InstallationAuthentication
~/project
# Authenticate with an API key
edge auth login

# Create a VM
edge compute create --name web-1 --size s-2vcpu-4gb --image ubuntu-24

# Upload to Storage
edge storage cp ./backup.tar.gz my-bucket/backups/

# Purge CDN cache
edge cdn purge my-deployment --path /images/*

# Add bot protection to a form
edge shield create --name "Contact form" --hostname example.com

REST API

The same API the console uses

Everything you can do in the console, you can do over HTTPS at https://edge.network/api. There's no cut-down surface for programmatic access.
  • Bearer authentication with ek_live_ API keys
  • Stable, machine-readable error codes
  • 600 requests a minute per IP, with X-RateLimit headers
  • Versioned with X-API-Version and at least six months' notice on deprecations
API referenceErrors
Create a VM
curl -X POST https://edge.network/api/compute/vms \
  -H "Authorization: Bearer ek_live_..." \
  -H "Content-Type: application/json" \
  -d '{
    "name": "web-01",
    "osTemplate": "ubuntu-24.04",
    "vcpu": 2,
    "ram": 4,
    "disk": 80,
    "region": "london"
  }'
Error response
{
  "error": "VM not found",
  "code": "NOT_FOUND"
}

Authentication

The right credential for each integration

API keys carry full account access, so keep one per integration and revoke individually. For AI agents and third-party tools, use an agent access code: per-product permissions and a monthly budget cap.
  • Keys start with ek_live_, with optional expiry
  • Manage keys programmatically at /api/account/keys
  • Agent codes start with ea_live_, scoped and revocable instantly
Auth guideAccess codes

Agent API and MCP

An API designed for how agents work

The official MCP server gives Cursor, Windsurf, Claude Desktop and any MCP client 22 native tools. Every call runs through the same Agent API, so scopes, budget caps and audit logging apply identically.
  • Discovery first. One call to GET /agent returns limits, budget, projects and a self-describing schema.
  • Dry run everything. Add X-Dry-Run: true to preview any mutation with a cost estimate.
  • Safe retries. Idempotency keys on every mutating endpoint.
  • tell_user on every response. A plain-English summary the agent can relay as is.
  • Self-signup. With no code set, the MCP server creates a free explore-tier account with a hard $0 cap.
Connect your toolAll 22 tools
mcp.json · local (stdio)
{
  "mcpServers": {
    "edge": {
      "command": "npx",
      "args": ["-y", "@edge-network/mcp"],
      "env": { "EDGE_AGENT_CODE": "ea_live_..." }
    }
  }
}
remote · streamable HTTP
POST https://edge.network/mcp
Content-Type: application/json
Authorization: Bearer ea_live_...

{ "jsonrpc": "2.0", "id": 1, "method": "tools/list" }

SDKs

Use the libraries you already know

Edge Storage works with the official AWS S3 SDKs. For agents, official wrappers expose the Edge tool set to LangChain, LlamaIndex, CrewAI and the Vercel AI SDK, fetched live so they never go stale.
S3 SDK examplesFramework wrappers
JavaScript · @aws-sdk/client-s3
import { S3Client } from '@aws-sdk/client-s3'

const client = new S3Client({
  endpoint: 'https://storage.edge.network',
  region: 'us-east-1',
  forcePathStyle: true,
  credentials: {
    accessKeyId: process.env.EDGE_ACCESS_KEY,
    secretAccessKey: process.env.EDGE_SECRET_KEY
  }
})
Python · boto3
import boto3
import os

client = boto3.client(
    's3',
    endpoint_url='https://storage.edge.network',
    aws_access_key_id=os.environ['EDGE_ACCESS_KEY'],
    aws_secret_access_key=os.environ['EDGE_SECRET_KEY'],
    region_name='us-east-1'
)
Python · LangChain
pip install "edge-network-agent[langchain]"

from edge_network_agent import EdgeAgentTools
from langchain.agents import create_react_agent

edge = EdgeAgentTools(agent_code="ea_live_...")
agent = create_react_agent(model, tools=edge.langchain_tools())
TypeScript · Vercel AI SDK
npm install @edge-network/ai-tools ai

import { generateText } from 'ai'
import { createEdgeTools, fetchStarterPrompt } from '@edge-network/ai-tools'

const tools = await createEdgeTools({ agentCode: 'ea_live_...' })

const { text } = await generateText({
  model,
  system: await fetchStarterPrompt(),
  tools,
  prompt: 'Deploy the ./dist folder as a static site and report the URL',
})

Machine-readable

Specs and discovery files at stable URLs

For tools, agents and integrations. Point your client at the URL and it gets structured answers, not HTML to scrape.

FAQ

Developer questions

Stuck on something? Get support.
Should I use an API key or an agent access code?
Use an API key (ek_live_) for your own scripts, the CLI and CI pipelines: it has full access to the account. Use an agent access code (ea_live_) for AI agents and third-party tools: you choose which products it can touch, whether it can deploy, manage or delete, and an optional soft or hard monthly budget cap.
Is there an official SDK?
The REST API is plain JSON over HTTPS, so any HTTP client works. For storage, use the official AWS S3 SDK for your language with the Edge endpoint. For agents, there's the edge-network-agent package on PyPI (LangChain, LlamaIndex, CrewAI) and @edge-network/ai-tools on npm for the Vercel AI SDK.
What are the rate limits?
The REST API allows 600 requests per minute per IP. Every response includes X-RateLimit-Limit and X-RateLimit-Remaining headers; going over returns 429 with a Retry-After header.
Can an agent create its own account?
Yes. Agents can sign up with an Ed25519 keypair and a small proof of work, with no card and no human in the loop. The account starts in a free explore tier with a hard $0 budget; a human claims it later to unlock paid resources.
How do you handle breaking changes?
The API is versioned with the X-API-Version header. Deprecations are signalled with Sunset headers and come with at least six months to migrate.

Build something on Edge

Free tiers on every product. Sign up in the console, or let your agent sign itself up: no card, no human in the loop.

Create free accountAgent self-signup

No card required to sign up.