Queues

Private network built for autonomous agents

One connector, one binary for your devices, servers, VPCs, and Workers on one private network, routed through Cloudflare's global network across 330+ cities.

Your private network across Region: Earth

Routes through the same global network across 330+ cities, handling 20% of Internet traffic. No relay servers. No degraded fallback paths.

Integrate with Workers

Bind your entire private network to Workers VPC with one line in wrangler.jsonc. Call any internal host with env.MESH.fetch().

Security built-in

Gateway policies, DNS filtering, DLP, and device posture checks built in

One private network for all your agents, services and clients

Background Pattern

Install the connector

One binary for laptops, Workers, or servers. No complex VPN setup, no relay servers. Just install and connect.

Background Pattern

Join the mesh

Your device gets a private IP on the mesh network automatically. Configure routes and policies from the dashboard.

Background Pattern

Routed through 330+ cities

Traffic flows through Cloudflare's global network, the same infrastructure handling 20% of Internet traffic.

Background Pattern

Reach any private service

Call databases, APIs, and internal hosts by private IP. Zero public endpoints exposed, with Zero Trust policies built in.

Background Pattern

Pull-consumers

Gain granular control over message consumption from anywhere on the public Internet by having your services explicitly pull messages and acknowledge them upon successful processing.

Background Pattern
Queues

Built for how agents actually work

Agents don't follow human workflows. They call APIs, query databases, and coordinate with other agents autonomously. Mesh gives them the private network they need.

View docs

Coding agents on private infra

Your local coding agent reaches private services over Mesh. No SSH tunnels, no port forwarding.

Workers calling private databases

Bind your Workers VPC to Mesh. Call internal Postgres, MySQL, or any TCP service from the edge.

Multi-agent coordination