pledgegateway

TypeScript SDK for PledgeGateway — Rust-native LLM gateway with prompt compression, caching, and cost tracking.

Install

npm install pledgegateway

Quick Start

import { PledgeGateway } from "pledgegateway";

const gateway = new PledgeGateway({
  baseUrl: process.env.PLEDGE_GATEWAY_URL || "http://localhost:8080",
  apiKey: process.env.PLEDGE_GATEWAY_API_KEY,
});

// Chat completion
const response = await gateway.chatCompletions({
  model: "gpt-4o",
  messages: [{ role: "user", content: "Hello!" }],
});

console.log(response.choices[0].message.content);

React Hooks

import { useChat, useStreamingChat } from "pledgegateway/react";

function Chat() {
  const { messages, input, setInput, send, isLoading } = useChat({
    baseUrl: "http://localhost:8080",
    model: "gpt-4o",
  });

  return (
    <div>
      {messages.map((m, i) => <p key={i}>{m.role}: {m.content}</p>)}
      <input value={input} onChange={(e) => setInput(e.target.value)} />
      <button onClick={send} disabled={isLoading}>Send</button>
    </div>
  );
}

Next.js Middleware

// middleware.ts
import { createMiddleware } from "pledgegateway/middleware";

export const middleware = createMiddleware({
  baseUrl: process.env.PLEDGE_GATEWAY_URL || "http://localhost:8080",
});

export const config = {
  matcher: ["/api/llm/:path*"],
};

API

PledgeGateway

• chatCompletions(req) — Chat completion
• chatCompletionsStream(req) — Streaming chat (async iterator)
• embeddings(req) — Embeddings
• stats() — Gateway statistics
• health() — Health check
• mcp(method, params) — MCP JSON-RPC call
• a2a(method, params) — A2A JSON-RPC call
• pledgestackStatus() — PledgeStack integration status
• pledgestackFlags() — Feature flags from PledgeStack

React Hooks

• useChat(options) — Chat with message history
• useStreamingChat(options) — Streaming chat with SSE
• useCompletion(options) — Single completion
• useGateway(config) — Gateway client instance
• useGatewayStats(config) — Live stats polling

License

MIT