Durable Streams
Deploy now
Deploy now
✨ Markdown

Durable Proxy ​

Use @durable-streams/proxy when you want to put durable, resumable streaming in front of existing AI streaming APIs.

The proxy forwards requests to an upstream provider, persists the streaming response into Durable Streams, and gives clients a durable read URL they can reconnect to.

Install ​

bash
pnpm add @durable-streams/proxy

Start a proxy server ​

typescript
import { createProxyServer } from "@durable-streams/proxy"

const server = await createProxyServer({
  port: 4440,
  durableStreamsUrl: "http://localhost:4441",
  jwtSecret: process.env.JWT_SECRET,
  allowlist: ["https://api.openai.com/**", "https://api.anthropic.com/**"],
})

console.log(`Proxy running at ${server.url}`)

Durable fetch ​

typescript
import { createDurableFetch } from "@durable-streams/proxy/client"

const durableFetch = createDurableFetch({
  proxyUrl: "https://my-proxy.example.com/v1/proxy",
  proxyAuthorization: "service-secret",
  autoResume: true,
  storage: localStorage,
})

const response = await durableFetch(
  "https://api.openai.com/v1/chat/completions",
  {
    method: "POST",
    headers: {
      "Content-Type": "application/json",
      Authorization: "Bearer sk-...",
    },
    body: JSON.stringify({
      model: "gpt-4",
      messages: [{ role: "user", content: "Hello" }],
      stream: true,
    }),
    requestId: "conversation-123",
  }
)

requestId lets the client resume the same durable stream across refreshes and reconnects.

When to use it ​

  • Use the proxy when you already have an SSE or streaming AI endpoint and want resumability without changing the upstream protocol.
  • Use Vercel AI SDK when you are integrating directly with Vercel AI SDK.
  • Use TanStack AI when you want the TanStack AI transport adapter.

More ​