Human-in-the-Loop Agent Task Management for Claude Code
Back to Glossary

SSE (Server-Sent Events)

Server-Sent Events (SSE) is a web technology that enables a server to push real-time, one-directional updates to a client over a persistent HTTP connection. Unlike traditional HTTP where the client must poll for updates, SSE maintains an open connection through which the server can stream data as it becomes available.

How SSE Works

`

Client Server

│ │

│── GET /events (EventSource) ──▶│

│ │

│◀── data: {"type":"task"} ─────│ (immediate push)

│ │

│◀── data: {"type":"message"} ──│ (seconds later)

│ │

│ [connection stays open] │

`

  1. The client opens a persistent HTTP connection to the server's SSE endpoint
  2. The server sends events as they occur — no polling required
  3. The connection auto-reconnects if dropped

SSE vs. WebSockets vs. Polling

Method Direction Protocol Use Case
Polling Client pulls HTTP Simple, high-latency OK
SSE Server pushes HTTP Real-time server-to-client
WebSockets Bidirectional WS Full-duplex real-time

SSE is simpler than WebSockets and works over standard HTTP/2 — making it ideal for server-to-agent event streaming.

SSE in MCP

The Model Context Protocol supports SSE as one of its transport options. When an MCP server like AgentRQ is accessed remotely, it uses HTTP for tool calls and SSE for streaming responses and real-time events.

SSE in AgentRQ

AgentRQ uses SSE to deliver real-time notifications and messages to connected clients. When a Claude Code agent calls the reply MCP tool, AgentRQ pushes the message to your task board via SSE — appearing instantly without any refresh or polling on your end.

Related Terms

Start Free