lightnode-sdk 0.7.14 → 0.7.16

package/dist/add.d.ts

@@ -48,6 +48,34 @@ export declare function addChat(opts?: AddOpts): {
     template: Template;
     network: Network;
 };
+/**
+ * `lightnode add chat-web3` - the user-pays counterpart to addChat.
+ *
+ * Architecture:
+ *   - Each visitor's own wallet signs SIWE + createSession per turn.
+ *   - The dev's app holds zero funds; cost is borne by each user (0.02 LCAI
+ *     per llama3-8b turn on mainnet).
+ *   - No backend, no PRIVATE_KEY, no server-side hot wallet.
+ *
+ * Fit:
+ *   - Web3 dApps where users already have a wallet (NFT, meme coin, on-chain
+ *     games, LightChallenge-style challenge platforms).
+ *   - For SaaS chatbots where users do NOT have a wallet, use `add chat`
+ *     instead (dev pays, server-side route).
+ */
+export declare function addChatWeb3(opts?: AddOpts): {
+    written: WrittenFile[];
+    install: string;
+    template: Template;
+    network: Network;
+    needsWagmi: boolean;
+};
+export declare function addWagmiSetup(opts?: AddOpts): {
+    written: WrittenFile[];
+    install: string;
+    template: Template;
+    network: Network;
+};
 export declare function addJudge(opts?: AddOpts): {
     written: WrittenFile[];
     install: string;

package/dist/add.js

@@ -31,6 +31,75 @@ function detectTemplate(cwd) {
         return "hono";
     return "node";
 }
+const HOSTING_GUIDE = `# Hosting LightChain AI inference - your options
+
+A single LightChain mainnet inference takes **60-90 seconds** under normal
+load (the workers do the model run, attest the result on-chain, and return
+the answer). If your server-side route has a function timeout shorter than
+that, every call will fail with a generic timeout error.
+
+This file is a quick reference for picking a host that actually finishes
+the request. If you used \`lightnode add chat-web3\`, ignore everything
+below - that path has NO server-side route to worry about.
+
+## TL;DR
+
+| Need                                  | Use                            |
+|---------------------------------------|--------------------------------|
+| Zero infra, users pay (Web3 dApp)     | \`lightnode add chat-web3\`      |
+| Cheapest server-side host that works  | Railway / Fly.io / Render      |
+| Already on Vercel and need it to fit  | Vercel Pro (60s) + streaming   |
+| You self-host already (Docker, VPS)   | Anywhere - no timeout to fight |
+
+## The detailed table
+
+| Host                | Free / Hobby timeout       | Paid timeout                | Verdict |
+|---------------------|----------------------------|-----------------------------|---------|
+| Vercel Hobby (free) | **10 seconds**             | -                           | **DOES NOT WORK** for mainnet inference. Every call times out. |
+| Vercel Pro ($20/mo) | -                          | 60s default, up to 800s with \`maxDuration\` config | Works if calls finish under 60s; 70-80s calls cut it close. Stream tokens to keep the connection warm. |
+| Cloudflare Workers  | 30s CPU                    | unbounded with Durable Objects | Free tier 30s is too tight; Workers + DO works but is complex. |
+| Railway             | none                       | none                        | **Great fit.** $5/mo for the smallest container, no timeout. Deploy a plain Node server or Next.js. |
+| Fly.io              | none                       | none                        | **Great fit.** Free tier covers small apps, scales to paid. |
+| Render              | none                       | none                        | **Great fit.** $7/mo for the smallest web service. |
+| Netlify Functions   | 26s sync / 15min background | 26s sync / 15min background | Use background functions for inference; sync functions time out. |
+| AWS Lambda          | 15 min                     | 15 min                      | Works, but WebSocket setup for streaming is involved. |
+| Self-host (Docker)  | no limit                   | no limit                    | **Works anywhere.** ECS, Cloud Run, your own VPS, fine. |
+
+## What I recommend
+
+1. **Building a Web3 app where users have wallets?** Switch to \`lightnode add
+   chat-web3\`. Each user signs and pays from their own wallet. You host the
+   static page (Vercel/Netlify/Cloudflare Pages free tier all work). No timeout
+   to fight because there is no server-side inference route.
+
+2. **Building a SaaS chatbot where users do NOT have wallets?** Deploy your
+   inference route on Railway or Fly. Both have no function timeout, scale
+   to traffic, and cost ~$5/mo for small apps. Cheaper than Vercel Pro.
+
+3. **Already committed to Vercel?** Upgrade to Pro and add streaming. The
+   chat template generated by \`lightnode add chat\` already uses streaming
+   tokens, so the connection stays warm while the model runs.
+
+4. **Have a Docker setup already?** Run the inference route in your existing
+   container. No timeout, no per-call cost beyond your existing infra.
+
+## Why the request is slow at all
+
+LightChain inference is not a synchronous LLM call. Each request:
+1. Negotiates an ECDH-encrypted session with a worker (off-chain).
+2. Sends a \`createSession\` tx on-chain (one confirmation).
+3. Uploads the encrypted prompt blob to the gateway.
+4. Sends a \`submitJob\` tx on-chain (one confirmation).
+5. Waits for the worker to run inference + post the encrypted result.
+6. Decrypts the result client-side.
+
+Steps 2, 4, and 5 are the slow part - each waits for a block confirmation
+and worker pickup. The protocol's verifiable-AI guarantee comes from doing
+all of this on-chain instead of just hitting an OpenAI-style API, which is
+the reason the call takes 60-90s instead of 1-2s. There is no way to
+shortcut this on the SDK side; the host just has to allow long-running
+functions.
+`;
 const ENV_EXAMPLE = (net) => `# Funded private key. Testnet works free (faucet at https://lightfaucet.ai).
 PRIVATE_KEY=0x0000000000000000000000000000000000000000000000000000000000000000
 
@@ -182,6 +251,7 @@ export function addInference(opts = {}) {
     const written = [];
     if (template === "nextjs-api") {
         written.push(writeFile(path.join(cwd, "app/api/inference/route.ts"), NEXTJS_ROUTE, force));
+        written.push(writeFile(path.join(cwd, "LIGHTNODE-HOSTING.md"), HOSTING_GUIDE, force));
     }
     else if (template === "hono") {
         written.push(writeFile(path.join(cwd, "lightchain-inference.ts"), HONO_HANDLER, force));
@@ -590,20 +660,47 @@ export default function Chat() {
     setTurns(next);
     setDraft("");
     setBusy(true);
+    // Reserve the assistant bubble immediately so tokens can stream into it.
+    setTurns([...next, { role: "assistant", text: "" }]);
     try {
-      // Concatenate prior turns so the model has the conversation context.
       const prompt =
         next.map((t) => (t.role === "user" ? \`User: \${t.text}\` : \`Assistant: \${t.text}\`)).join("\\n") +
         "\\nAssistant:";
       const r = await fetch("/api/inference", {
         method: "POST",
         headers: { "Content-Type": "application/json" },
-        body: JSON.stringify({ prompt }),
-      }).then((r) => r.json()) as { answer?: string; txs?: Record<string, string>; error?: string };
-      if (!r.answer) throw new Error(r.error ?? "no answer");
-      setTurns([...next, { role: "assistant", text: r.answer, txs: r.txs }]);
+        body: JSON.stringify({ prompt, stream: true }),
+      });
+      if (!r.ok || !r.body) {
+        const err = await r.text().catch(() => "");
+        throw new Error(\`route returned \${r.status}: \${err.slice(0, 200)}\`);
+      }
+      const reader = r.body.getReader();
+      const decoder = new TextDecoder();
+      let assembled = "";
+      let txs: Record<string, string> | undefined;
+      while (true) {
+        const { done, value } = await reader.read();
+        if (done) break;
+        const text = decoder.decode(value, { stream: true });
+        // The route ends the stream with one line of JSON on a new \\u0000
+        // delimiter (a NUL byte we put between the streamed prose and the
+        // metadata). Anything before is body; anything after is metadata.
+        const nulIdx = text.indexOf("\\u0000");
+        if (nulIdx === -1) {
+          assembled += text;
+          setTurns((prev) => prev.map((t, i) => i === prev.length - 1 ? { ...t, text: assembled } : t));
+        } else {
+          assembled += text.slice(0, nulIdx);
+          try { txs = JSON.parse(text.slice(nulIdx + 1)) as Record<string, string>; } catch { /* ignore */ }
+          setTurns((prev) => prev.map((t, i) => i === prev.length - 1 ? { ...t, text: assembled, txs } : t));
+        }
+      }
     } catch (e) {
-      setTurns([...next, { role: "assistant", text: \`(error: \${(e as Error).message})\` }]);
+      // Replace the empty assistant bubble with the error.
+      setTurns((prev) => prev.map((t, i) =>
+        i === prev.length - 1 ? { ...t, text: \`(error: \${(e as Error).message})\` } : t
+      ));
     } finally {
       setBusy(false);
     }
@@ -664,13 +761,300 @@ export default function Chat() {
         </button>
       </div>
       <p style={{ marginTop: 16, fontSize: 12, color: "#888" }}>
-        Make sure <code>/api/inference</code> is mounted - run <code>npx lightnode add inference</code> if you have
-        not. The route signs every call server-side with the PRIVATE_KEY in your .env.
+        Your server signs every call with the PRIVATE_KEY in <code>.env</code>. Cost per turn is paid from that wallet.
+        See <code>LIGHTNODE-HOSTING.md</code> for picking a host that handles 60-90s function calls.
       </p>
     </main>
   );
 }
 `;
+/**
+ * Streaming inference route, paired with NEXTJS_CHAT_PAGE.
+ *
+ * Why a second route: the plain NEXTJS_ROUTE returns JSON in one shot,
+ * which is fine for one-off calls but means the chat UI shows nothing for
+ * 60-90s. This streams decrypted chunks as they arrive (via
+ * runInferenceStream), so the user sees tokens land live and the host
+ * keeps the connection warm. After the last chunk we append a NUL byte
+ * (\\u0000) and a JSON line with the tx hashes so the client can render
+ * the on-chain receipt under the assistant bubble.
+ */
+const NEXTJS_INFERENCE_STREAM_ROUTE = `// app/api/inference/route.ts
+// Generated by 'lightnode add chat'. Streaming inference route - tokens
+// arrive live as the model produces them, then a NUL byte separates them
+// from a final JSON line with the on-chain tx hashes.
+//
+// Pass { stream: true } in the body to get streaming. Without that flag
+// the same route still returns one-shot JSON, which is fine for non-chat
+// integrations.
+//
+// Mainnet 8b inference takes 60-90s. Vercel Hobby caps function execution
+// at 10s and will time out. See LIGHTNODE-HOSTING.md for hosts that work.
+import { NextResponse } from "next/server";
+import { runInferenceWithKey, runInferenceStream } from "lightnode-sdk";
+
+export const runtime = "nodejs";
+export const dynamic = "force-dynamic";
+export const maxDuration = 120;
+
+const NETWORK = (process.env.NETWORK ?? "testnet") as "mainnet" | "testnet";
+const MODEL = process.env.MODEL ?? "llama3-8b";
+
+export async function POST(req: Request) {
+  if (!process.env.PRIVATE_KEY?.startsWith("0x")) {
+    return NextResponse.json({ error: "PRIVATE_KEY not configured" }, { status: 500 });
+  }
+  const body = (await req.json().catch(() => ({}))) as { prompt?: string; system?: string; stream?: boolean };
+  const prompt = body.prompt?.trim();
+  if (!prompt) return NextResponse.json({ error: "prompt is required" }, { status: 400 });
+
+  const args = {
+    network: NETWORK,
+    privateKey: process.env.PRIVATE_KEY as \`0x\${string}\`,
+    model: MODEL,
+    system: body.system?.trim() || undefined,
+    prompt,
+  };
+
+  // One-shot JSON path: same shape as the original 'add inference' route.
+  if (!body.stream) {
+    try {
+      const { answer, worker, txs, jobId } = await runInferenceWithKey(args);
+      return NextResponse.json({
+        answer, worker, jobId: jobId.toString(),
+        txs: { createSession: txs.createSession, submitJob: txs.submitJob, jobCompleted: txs.jobCompleted },
+      });
+    } catch (e) {
+      return NextResponse.json({ error: (e as Error).message }, { status: 500 });
+    }
+  }
+
+  // Streaming path: tokens, NUL byte, then tx metadata as JSON.
+  const stream = runInferenceStream(args);
+  const encoder = new TextEncoder();
+  const body$ = new ReadableStream({
+    async start(controller) {
+      try {
+        for await (const chunk of stream) controller.enqueue(encoder.encode(chunk));
+        const { worker, txs, jobId } = await stream.done;
+        const meta = JSON.stringify({
+          worker, jobId: jobId.toString(),
+          createSession: txs.createSession, submitJob: txs.submitJob, jobCompleted: txs.jobCompleted,
+        });
+        controller.enqueue(encoder.encode("\\u0000" + meta));
+        controller.close();
+      } catch (e) {
+        controller.enqueue(encoder.encode(\`\\u0000{"error":\${JSON.stringify((e as Error).message)}}\`));
+        controller.close();
+      }
+    },
+  });
+  return new Response(body$, {
+    headers: { "Content-Type": "text/plain; charset=utf-8", "Cache-Control": "no-store" },
+  });
+}
+`;
+const NEXTJS_CHAT_WEB3_PAGE = `// app/chat-web3/page.tsx
+// Generated by 'lightnode add chat-web3'. User-pays chat: each visitor's own
+// wallet signs SIWE + createSession per turn. Your app holds zero funds.
+//
+// Prereqs:
+//   - wagmi configured in your app (https://wagmi.sh/react/getting-started)
+//   - the connected wallet has LCAI on the LightChain network it is on
+//     (mainnet 9200 or testnet 8200). Mainnet llama3-8b costs 0.02 LCAI per
+//     turn plus a small gas amount.
+"use client";
+
+import { useEffect, useState } from "react";
+import { useAccount, useWalletClient, usePublicClient } from "wagmi";
+import { siweSignIn, GatewayClient, runInference, estimateJobFee, NETWORKS } from "lightnode-sdk";
+
+type Turn = {
+  role: "user" | "assistant";
+  text: string;
+  worker?: string | null;
+  jobId?: string | null;
+  submitTx?: \`0x\${string}\` | null;
+  jobCompletedTx?: \`0x\${string}\` | null;
+};
+
+const MODEL = "llama3-8b";
+
+export default function ChatWeb3() {
+  const { address, chain } = useAccount();
+  const network: "mainnet" | "testnet" | null =
+    chain?.id === 9200 ? "mainnet" : chain?.id === 8200 ? "testnet" : null;
+  const { data: walletClient } = useWalletClient({ chainId: chain?.id });
+  const publicClient = usePublicClient({ chainId: chain?.id });
+
+  const [turns, setTurns] = useState<Turn[]>([]);
+  const [input, setInput] = useState("");
+  const [busy, setBusy] = useState(false);
+  const [busyStage, setBusyStage] = useState("");
+  const [err, setErr] = useState<string | null>(null);
+  const [feeLcai, setFeeLcai] = useState<number | null>(null);
+
+  // Read the on-chain fee for the connected network so we can show the
+  // visitor the real cost per turn before they click Send.
+  useEffect(() => {
+    if (!network) { setFeeLcai(null); return; }
+    let cancelled = false;
+    estimateJobFee(NETWORKS[network], MODEL).then(
+      (fee) => { if (!cancelled) setFeeLcai(fee); },
+      () => { if (!cancelled) setFeeLcai(null); },
+    );
+    return () => { cancelled = true; };
+  }, [network]);
+
+  /** Build a single prompt from history + new user input. */
+  function composePrompt(history: Turn[], next: string, system: string): string {
+    const lines: string[] = [];
+    for (const t of history) {
+      lines.push(\`\${t.role === "user" ? "User" : "Assistant"}: \${t.text}\`);
+    }
+    lines.push(\`User: \${next}\`);
+    lines.push("Assistant:");
+    return system ? \`\${system}\\n\\n\${lines.join("\\n\\n")}\` : lines.join("\\n\\n");
+  }
+
+  async function send() {
+    if (!walletClient || !publicClient || !address || !network) {
+      setErr("Connect a wallet on LightChain mainnet (9200) or testnet (8200) first.");
+      return;
+    }
+    const next = input.trim();
+    if (!next) return;
+    setBusy(true);
+    setErr(null);
+    const history = [...turns];
+    // Optimistic user bubble so it appears immediately.
+    setTurns([...history, { role: "user", text: next }]);
+    setInput("");
+    try {
+      const system = "You are a concise assistant. Reply in one or two short sentences.";
+      const prompt = composePrompt(history, next, system);
+
+      setBusyStage("Sign in with your wallet (SIWE)...");
+      const session = await siweSignIn(walletClient as unknown as Parameters<typeof siweSignIn>[0], network);
+
+      setBusyStage("Approve the createSession transaction in your wallet...");
+      const gateway = new GatewayClient({ network, bearer: session.bearer });
+      const result = await runInference({
+        prompt,
+        gateway,
+        wallet: walletClient as unknown as Parameters<typeof runInference>[0]["wallet"],
+        publicClient: publicClient as unknown as Parameters<typeof runInference>[0]["publicClient"],
+        network: NETWORKS[network],
+        model: MODEL,
+        jobCompletedTimeoutMs: 120_000,
+        maxRetries: 1,
+      });
+
+      setTurns([...history, { role: "user", text: next }, {
+        role: "assistant",
+        text: result.answer,
+        worker: result.worker,
+        jobId: result.jobId?.toString() ?? null,
+        submitTx: result.txs?.submitJob ?? null,
+        jobCompletedTx: result.txs?.jobCompleted ?? null,
+      }]);
+    } catch (e) {
+      // Roll back the optimistic user bubble so the visitor can retry.
+      setTurns(history);
+      setInput(next);
+      const msg = (e as Error).message ?? "chat failed";
+      setErr(
+        /user rejected|user denied|reject/i.test(msg)
+          ? "You rejected the wallet popup. Try again."
+          : /insufficient funds|insufficient balance/i.test(msg)
+            ? \`Your wallet has no \${network === "mainnet" ? "LCAI" : "testnet LCAI"}. Top it up before sending.\`
+            : msg.split("\\n")[0],
+      );
+    } finally {
+      setBusy(false);
+      setBusyStage("");
+    }
+  }
+
+  return (
+    <main style={{ maxWidth: 720, margin: "32px auto", padding: 16, fontFamily: "system-ui" }}>
+      <h1>Chat (user-pays)</h1>
+      <p style={{ color: "#666" }}>
+        Each turn signs one createSession transaction from your wallet on{" "}
+        <code>{network ?? "(connect a wallet)"}</code>. Fee:{" "}
+        <code>{feeLcai != null ? \`\${feeLcai} LCAI\` : "(fetching)"}</code> per turn plus a small gas amount.
+      </p>
+      {!address && (
+        <div style={{ border: "1px solid #ddd", borderRadius: 8, padding: 12, margin: "12px 0" }}>
+          Connect a wallet to start chatting. (Use whichever connector your app exposes - e.g. RainbowKit,
+          ConnectKit, Reown AppKit, or wagmi&apos;s useConnect directly.)
+        </div>
+      )}
+
+      <div style={{ display: "flex", flexDirection: "column", gap: 8, margin: "16px 0" }}>
+        {turns.map((t, i) => (
+          <div
+            key={i}
+            style={{
+              alignSelf: t.role === "user" ? "flex-end" : "flex-start",
+              maxWidth: "85%",
+              borderRadius: 12,
+              padding: "8px 12px",
+              background: t.role === "user" ? "#e9e7ff" : "#f5f5f7",
+            }}
+          >
+            <div style={{ whiteSpace: "pre-wrap", wordBreak: "break-word" }}>{t.text}</div>
+            {t.role === "assistant" && t.submitTx ? (
+              <div style={{ marginTop: 6, fontSize: 11, color: "#666", display: "flex", gap: 8, flexWrap: "wrap" }}>
+                {t.worker && (
+                  <a href={\`https://\${network}.lightscan.app/address/\${t.worker}\`} target="_blank" rel="noopener noreferrer">
+                    worker
+                  </a>
+                )}
+                {t.jobId && <span>job #{t.jobId}</span>}
+                {t.submitTx && (
+                  <a href={\`https://\${network}.lightscan.app/tx/\${t.submitTx}\`} target="_blank" rel="noopener noreferrer">
+                    submitJob
+                  </a>
+                )}
+                {t.jobCompletedTx && (
+                  <a href={\`https://\${network}.lightscan.app/tx/\${t.jobCompletedTx}\`} target="_blank" rel="noopener noreferrer">
+                    completed
+                  </a>
+                )}
+              </div>
+            ) : null}
+          </div>
+        ))}
+      </div>
+
+      <textarea
+        value={input}
+        onChange={(e) => setInput(e.target.value)}
+        onKeyDown={(e) => {
+          if (e.key === "Enter" && (e.metaKey || e.ctrlKey)) { e.preventDefault(); if (!busy && input.trim()) send(); }
+        }}
+        rows={2}
+        placeholder={turns.length === 0 ? "Say hello (cmd+enter to send)" : "Reply..."}
+        style={{ width: "100%", padding: 8, fontFamily: "inherit" }}
+      />
+      <button
+        type="button"
+        onClick={() => send()}
+        disabled={busy || !input.trim() || !address || !network}
+        style={{ marginTop: 8, padding: "8px 16px" }}
+      >
+        {busy ? (busyStage || "Sending...") : (turns.length === 0 ? "Send first message" : "Send")}
+      </button>
+      {err && (
+        <p style={{ marginTop: 8, padding: "8px 12px", border: "1px solid #f5c2c7", background: "#f8d7da", color: "#842029", borderRadius: 6 }}>
+          {err}
+        </p>
+      )}
+    </main>
+  );
+}
+`;
 const NODE_CHAT_REPL = `// chat-repl.ts
 // Generated by 'lightnode add chat'. Interactive chat REPL in your terminal.
 //   npm install lightnode-sdk viem ws
@@ -900,7 +1284,12 @@ export function addChat(opts = {}) {
     const force = !!opts.force;
     const written = [];
     if (template === "nextjs-api") {
+        // 'add chat' is now self-contained: it writes BOTH the chat page AND
+        // the streaming inference route. Previously the user had to remember
+        // to also run 'add inference' separately, which was easy to miss.
         written.push(writeFile(path.join(cwd, "app/chat/page.tsx"), NEXTJS_CHAT_PAGE, force));
+        written.push(writeFile(path.join(cwd, "app/api/inference/route.ts"), NEXTJS_INFERENCE_STREAM_ROUTE, force));
+        written.push(writeFile(path.join(cwd, "LIGHTNODE-HOSTING.md"), HOSTING_GUIDE, force));
     }
     else {
         written.push(writeFile(path.join(cwd, "chat-repl.ts"), NODE_CHAT_REPL, force));
@@ -908,6 +1297,191 @@ export function addChat(opts = {}) {
     written.push(writeFile(path.join(cwd, ".env.example"), ENV_EXAMPLE(network), force));
     return { written, install: `npm install ${depsNeeded(template).join(" ")}`, template, network };
 }
+/**
+ * `lightnode add chat-web3` - the user-pays counterpart to addChat.
+ *
+ * Architecture:
+ *   - Each visitor's own wallet signs SIWE + createSession per turn.
+ *   - The dev's app holds zero funds; cost is borne by each user (0.02 LCAI
+ *     per llama3-8b turn on mainnet).
+ *   - No backend, no PRIVATE_KEY, no server-side hot wallet.
+ *
+ * Fit:
+ *   - Web3 dApps where users already have a wallet (NFT, meme coin, on-chain
+ *     games, LightChallenge-style challenge platforms).
+ *   - For SaaS chatbots where users do NOT have a wallet, use `add chat`
+ *     instead (dev pays, server-side route).
+ */
+export function addChatWeb3(opts = {}) {
+    const cwd = opts.cwd ?? process.cwd();
+    const network = opts.network ?? "mainnet";
+    // chat-web3 is browser-only. If the project is not Next.js (or another
+    // React framework we'd detect), fall back to nextjs-api anyway and warn
+    // the user in the CLI's next-steps that the file expects a Next.js
+    // setup with wagmi.
+    const detected = detectTemplate(cwd);
+    const template = opts.template && opts.template !== "auto" ? opts.template : detected;
+    const force = !!opts.force;
+    const written = [];
+    // Detect whether the project already has wagmi; if not, the next-steps
+    // output prints the install line.
+    const pkg = readPackageJson(cwd);
+    const deps = { ...(pkg?.dependencies ?? {}), ...(pkg?.devDependencies ?? {}) };
+    const hasWagmi = Boolean(deps["wagmi"]);
+    written.push(writeFile(path.join(cwd, "app/chat-web3/page.tsx"), NEXTJS_CHAT_WEB3_PAGE, force));
+    return {
+        written,
+        install: `npm install lightnode-sdk viem` + (hasWagmi ? "" : " wagmi @tanstack/react-query"),
+        template,
+        network,
+        needsWagmi: !hasWagmi,
+    };
+}
+const WAGMI_CONFIG_FILE = `// lib/wagmi.ts
+// Generated by 'lightnode add wagmi-setup'. Minimal wagmi setup for
+// LightChain mainnet (9200) + testnet (8200). Use this as a starting
+// point; swap in RainbowKit / Reown AppKit / ConnectKit if you want a
+// richer connect UI.
+import { createConfig, http } from "wagmi";
+import { injected } from "wagmi/connectors";
+import type { Chain } from "viem";
+
+export const lightchainMainnet: Chain = {
+  id: 9200,
+  name: "LightChain",
+  nativeCurrency: { name: "LCAI", symbol: "LCAI", decimals: 18 },
+  rpcUrls: { default: { http: ["https://rpc.mainnet.lightchain.ai"] } },
+  blockExplorers: { default: { name: "Lightscan", url: "https://mainnet.lightscan.app" } },
+};
+
+export const lightchainTestnet: Chain = {
+  id: 8200,
+  name: "LightChain Testnet",
+  nativeCurrency: { name: "LCAI", symbol: "LCAI", decimals: 18 },
+  rpcUrls: { default: { http: ["https://rpc.testnet.lightchain.ai"] } },
+  blockExplorers: { default: { name: "Lightscan", url: "https://testnet.lightscan.app" } },
+};
+
+export const wagmiConfig = createConfig({
+  chains: [lightchainMainnet, lightchainTestnet],
+  // \`injected\` covers MetaMask, Rabby, OKX, Phantom EVM, and any browser
+  // wallet that follows EIP-1193. Add walletConnect() / coinbaseWallet()
+  // here if you want explicit support for those.
+  connectors: [injected()],
+  transports: {
+    [lightchainMainnet.id]: http(),
+    [lightchainTestnet.id]: http(),
+  },
+});
+`;
+const WAGMI_PROVIDERS_FILE = `// app/providers.tsx
+// Generated by 'lightnode add wagmi-setup'. Wraps the app with the
+// wagmi + react-query providers needed for any wagmi hook to work.
+//
+// Import this from your root layout:
+//
+//   // app/layout.tsx
+//   import { Providers } from "./providers";
+//   export default function RootLayout({ children }: { children: React.ReactNode }) {
+//     return (
+//       <html lang="en">
+//         <body><Providers>{children}</Providers></body>
+//       </html>
+//     );
+//   }
+"use client";
+
+import { WagmiProvider } from "wagmi";
+import { QueryClient, QueryClientProvider } from "@tanstack/react-query";
+import { wagmiConfig } from "@/lib/wagmi";
+import type { ReactNode } from "react";
+
+const queryClient = new QueryClient();
+
+export function Providers({ children }: { children: ReactNode }) {
+  return (
+    <WagmiProvider config={wagmiConfig}>
+      <QueryClientProvider client={queryClient}>
+        {children}
+      </QueryClientProvider>
+    </WagmiProvider>
+  );
+}
+`;
+const WAGMI_CONNECT_BUTTON = `// components/connect-button.tsx
+// Generated by 'lightnode add wagmi-setup'. Minimal Connect/Disconnect
+// button using wagmi's useConnect / useAccount hooks. Swap in
+// RainbowKit's ConnectButton or Reown's <w3m-button /> for a richer UI.
+"use client";
+
+import { useAccount, useConnect, useDisconnect, useSwitchChain } from "wagmi";
+
+const LIGHTCHAIN_IDS = new Set<number>([9200, 8200]);
+
+function shortAddress(addr: string): string {
+  return \`\${addr.slice(0, 6)}...\${addr.slice(-4)}\`;
+}
+
+export function ConnectButton() {
+  const { address, chain, isConnected } = useAccount();
+  const { connect, connectors, isPending } = useConnect();
+  const { disconnect } = useDisconnect();
+  const { switchChain, isPending: switching } = useSwitchChain();
+
+  if (!isConnected) {
+    const connector = connectors[0];
+    return (
+      <button
+        type="button"
+        onClick={() => connect({ connector })}
+        disabled={isPending}
+        style={{ padding: "8px 16px", borderRadius: 8, cursor: "pointer" }}
+      >
+        {isPending ? "Connecting..." : "Connect wallet"}
+      </button>
+    );
+  }
+
+  if (chain && !LIGHTCHAIN_IDS.has(chain.id)) {
+    return (
+      <button
+        type="button"
+        onClick={() => switchChain({ chainId: 9200 })}
+        disabled={switching}
+        style={{ padding: "8px 16px", borderRadius: 8, background: "#fee", cursor: "pointer" }}
+      >
+        {switching ? "Switching..." : "Switch to LightChain"}
+      </button>
+    );
+  }
+
+  return (
+    <button
+      type="button"
+      onClick={() => disconnect()}
+      style={{ padding: "8px 16px", borderRadius: 8, cursor: "pointer", fontFamily: "monospace" }}
+    >
+      {address ? shortAddress(address) : "(unknown)"} ({chain?.name}) - disconnect
+    </button>
+  );
+}
+`;
+export function addWagmiSetup(opts = {}) {
+    const cwd = opts.cwd ?? process.cwd();
+    const network = opts.network ?? "mainnet";
+    const template = opts.template && opts.template !== "auto" ? opts.template : detectTemplate(cwd);
+    const force = !!opts.force;
+    const written = [];
+    written.push(writeFile(path.join(cwd, "lib/wagmi.ts"), WAGMI_CONFIG_FILE, force));
+    written.push(writeFile(path.join(cwd, "app/providers.tsx"), WAGMI_PROVIDERS_FILE, force));
+    written.push(writeFile(path.join(cwd, "components/connect-button.tsx"), WAGMI_CONNECT_BUTTON, force));
+    return {
+        written,
+        install: `npm install wagmi viem @tanstack/react-query`,
+        template,
+        network,
+    };
+}
 const NEXTJS_JUDGE_ROUTE = `// app/api/judge/route.ts
 // Generated by 'lightnode add judge'. See https://lightnode.app/build
 // The LightChallenge-style evaluator: post evidence + criteria, get a
@@ -1041,6 +1615,7 @@ export function addJudge(opts = {}) {
     const written = [];
     if (template === "nextjs-api") {
         written.push(writeFile(path.join(cwd, "app/api/judge/route.ts"), NEXTJS_JUDGE_ROUTE, force));
+        written.push(writeFile(path.join(cwd, "LIGHTNODE-HOSTING.md"), HOSTING_GUIDE, force));
     }
     else {
         written.push(writeFile(path.join(cwd, "judge.ts"), NODE_JUDGE_SCRIPT, force));

package/dist/cli.js

@@ -1,6 +1,6 @@
 #!/usr/bin/env node
 import { LightNode, modelStatsCsv, workerStatsCsv, workerJobsCsv, runInferenceWithKey, runInferenceBatch, Agent, isStalledWorker, workerPreflight, workerWatch, WorkerOperator, isWorkerOpError, BRIDGE_ROUTE, DAO, DAO_ADDRESSES } from "./index.js";
-import { addInference, addAnalyticsDashboard, addNftMint, addChat, addAgent, addJudge } from "./add.js";
+import { addInference, addAnalyticsDashboard, addNftMint, addChat, addChatWeb3, addAgent, addJudge, addWagmiSetup } from "./add.js";
 import { createPublicClient, createWalletClient, http, parseEther } from "viem";
 import { privateKeyToAccount, generatePrivateKey } from "viem/accounts";
 function flag(name) {
@@ -462,7 +462,7 @@ async function main() {
             const template = flag("--template") ?? "auto";
             const force = process.argv.includes("--force");
             const network = (net === "mainnet" ? "mainnet" : "testnet");
-            const known = ["inference", "chat", "agent", "judge", "analytics-dashboard", "nft-mint-with-inference"];
+            const known = ["inference", "chat", "chat-web3", "wagmi-setup", "agent", "judge", "analytics-dashboard", "nft-mint-with-inference"];
             if (!known.includes(sub ?? "")) {
                 die(`usage: lightnode add <${known.join("|")}> [--template auto|nextjs-api|hono|node] [--net testnet|mainnet] [--force]`);
             }
@@ -470,13 +470,17 @@ async function main() {
                 ? addAnalyticsDashboard({ template, network, force })
                 : sub === "nft-mint-with-inference"
                     ? addNftMint({ template, network, force })
-                    : sub === "chat"
-                        ? addChat({ template, network, force })
-                        : sub === "agent"
-                            ? addAgent({ template, network, force })
-                            : sub === "judge"
-                                ? addJudge({ template, network, force })
-                                : addInference({ template, network, force });
+                    : sub === "chat-web3"
+                        ? addChatWeb3({ template, network, force })
+                        : sub === "wagmi-setup"
+                            ? addWagmiSetup({ template, network, force })
+                            : sub === "chat"
+                                ? addChat({ template, network, force })
+                                : sub === "agent"
+                                    ? addAgent({ template, network, force })
+                                    : sub === "judge"
+                                        ? addJudge({ template, network, force })
+                                        : addInference({ template, network, force });
             console.log(`▶ add ${sub} (${result.template} template, default network ${result.network})`);
             for (const f of result.written) {
                 if (f.skipped)
@@ -491,7 +495,36 @@ async function main() {
             else {
                 console.log(`\nNext steps (these files were added to your CURRENT folder, not a new project):`);
                 console.log(`  1. ${result.install}`);
-                if (sub === "nft-mint-with-inference" || sub === "inference" || sub === "chat" || sub === "agent" || sub === "judge") {
+                if (sub === "wagmi-setup") {
+                    console.log(`  2. Import Providers in app/layout.tsx and wrap children:`);
+                    console.log(`     import { Providers } from "./providers";`);
+                    console.log(`     // <body><Providers>{children}</Providers></body>`);
+                    console.log(`  3. Drop <ConnectButton /> anywhere you want a connect UI:`);
+                    console.log(`     import { ConnectButton } from "@/components/connect-button";`);
+                    console.log(`  4. You can now use any wagmi hook (useAccount, useWalletClient, ...).`);
+                    console.log(`     Wallets on chains other than 9200/8200 will be prompted to switch.`);
+                }
+                else if (sub === "chat-web3") {
+                    // chat-web3 has no PRIVATE_KEY (each visitor pays their own way).
+                    const needsWagmi = result.needsWagmi;
+                    if (needsWagmi) {
+                        console.log(`  2. Get wagmi wired up with one command:`);
+                        console.log(`     npx lightnode add wagmi-setup`);
+                        console.log(`     (drops lib/wagmi.ts + app/providers.tsx + components/connect-button.tsx)`);
+                        console.log(`  3. Wrap your layout with <Providers> (see step 2 output) and drop`);
+                        console.log(`     <ConnectButton /> somewhere on the page.`);
+                        console.log(`  4. npm run dev, open /chat-web3, connect on chainId ${result.network === "mainnet" ? "9200" : "8200"}.`);
+                        console.log(`     Mainnet llama3-8b costs 0.02 LCAI per turn; testnet is free from https://lightfaucet.ai`);
+                    }
+                    else {
+                        console.log(`  2. npm run dev, open /chat-web3`);
+                        console.log(`  3. Connect a wallet on LightChain ${result.network === "mainnet" ? "mainnet (chainId 9200)" : "testnet (chainId 8200)"}.`);
+                        console.log(`     Mainnet llama3-8b costs 0.02 LCAI per turn; testnet is free from https://lightfaucet.ai`);
+                    }
+                    console.log(`\n  Note: chat-web3 has NO server-side route, so it scales infinitely on`);
+                    console.log(`  static hosting (Vercel/Netlify/Cloudflare Pages free tier all work).`);
+                }
+                else if (sub === "nft-mint-with-inference" || sub === "inference" || sub === "chat" || sub === "agent" || sub === "judge") {
                     console.log(`  2. cp .env.example .env  (and put a funded ${result.network} PRIVATE_KEY in it)`);
                     if (sub === "agent" && result.template === "nextjs-api") {
                         console.log(`  3. Set CRON_SECRET in your Vercel env vars + edit AGENT_TASK in .env`);
@@ -501,8 +534,8 @@ async function main() {
                         console.log(`  3. AGENT_INTERVAL_MS=3600000 npx tsx agent.ts   # or run under pm2/systemd`);
                     }
                     else if (sub === "chat" && result.template === "nextjs-api") {
-                        console.log(`  3. Make sure /api/inference is mounted too (run: npx lightnode add inference)`);
-                        console.log(`  4. npm run dev, open /chat`);
+                        console.log(`  3. npm run dev, open /chat`);
+                        console.log(`     (the chat page + /api/inference streaming route are both already wired up)`);
                     }
                     else if (sub === "chat") {
                         console.log(`  3. npx tsx chat-repl.ts  (interactive terminal chat)`);
@@ -541,6 +574,13 @@ async function main() {
                         console.log(`  2. npx tsx lightnode-analytics.ts`);
                     }
                 }
+                // Hosting warning for server-side commands that ship LIGHTNODE-HOSTING.md.
+                if (result.template === "nextjs-api"
+                    && (sub === "inference" || sub === "chat" || sub === "judge")) {
+                    console.log(`\n  Hosting: a mainnet inference takes 60-90s. Vercel Hobby (free) times`);
+                    console.log(`  out at 10s; Vercel Pro at 60s. See LIGHTNODE-HOSTING.md (in this folder)`);
+                    console.log(`  for hosts that work - Railway / Fly / Render handle long calls fine.`);
+                }
                 if (result.network === "testnet") {
                     console.log(`\nNo wallet yet? Make one:  npx lightnode wallet new   then fund it free below.`);
                 }

package/dist/index.d.ts

@@ -134,7 +134,7 @@ export declare class LightNode {
  * (especially in registry-proxy environments like StackBlitz where lockfiles
  * may pin an older minor than the local install command suggests).
  */
-export declare const SDK_VERSION = "0.7.14";
+export declare const SDK_VERSION = "0.7.16";
 export { NETWORKS, WORKER_REGISTRY, REGISTRY_TOPICS, aggregateModelStats, aggregateWorkerStats, networkAnalytics, modelStatsCsv, workerStatsCsv, workerJobsCsv, fromWei, resolveJobTransactions, siweSignIn, siweChallenge, siweVerify, fetchWorkerModels, computeModelId as modelId, estimateJobFee, JOB_REGISTRY_CONSUMER_ABI, consumerGatewayUrl, consumerGatewayHost, GatewayClient, GatewayHttpError, prepareSession, submitPrompt, decryptResponse, generateEcdhKeyPair, crypto, runInference, runInferenceWithKey, runInferenceStream, Conversation, chat, runInferenceBatch, Agent, parseAgentOutput, workerPreflight, workerWatch, Bridge, BRIDGE_ROUTE, HYPERLANE_ROUTER_ABI, ERC20_ABI, addressToBytes32, quoteBridgeFee, bridgeableBalance, bridgeAllowance, approveBridge, bridgeTransfer, DAO, DAO_ADDRESSES, ProposalState, PROPOSAL_STATE_LABEL, VoteSupport, GOVERNOR_ABI, VOTES_ABI, OnchainModelRegistry, AIVM_MODEL_REGISTRY_ABI, BENCHMARK_REGISTRY_ABI, ModelStatus, MODEL_STATUS_LABEL, StalledWorkerError, OnChainRevertError, RelayTokenTimeoutError, GatewayAuthError, isStalledWorker, WorkerOperator, WORKER_REGISTRY_ABI, JOB_REGISTRY_OPERATOR_ABI, AI_CONFIG_ABI, JOB_STATE, decodeWorkerError, WorkerOpError, isWorkerOpError, };
 export type { BearerSource, GatewayClientOptions, SelectSessionResult, PrepareSessionResult, UploadBlobResult, SessionTokenResult } from "./gateway.js";
 export type { SessionPreparation, RunInferenceArgs, RunInferenceResult, RunInferenceWithKeyArgs, RunInferenceStreamResult } from "./inference.js";

package/dist/index.js

@@ -213,7 +213,7 @@ export class LightNode {
  * (especially in registry-proxy environments like StackBlitz where lockfiles
  * may pin an older minor than the local install command suggests).
  */
-export const SDK_VERSION = "0.7.14";
+export const SDK_VERSION = "0.7.16";
 export { NETWORKS, WORKER_REGISTRY, REGISTRY_TOPICS, aggregateModelStats, aggregateWorkerStats, networkAnalytics, modelStatsCsv, workerStatsCsv, workerJobsCsv, fromWei, 
 // v0.7.3 per-job transaction-hash resolver (lifts the upstream
 // subgraph's "block-only" Job entity to a deep-linkable Job + tx pair).

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "lightnode-sdk",
-  "version": "0.7.14",
+  "version": "0.7.16",
   "description": "Read-only TypeScript client for LightChain AI: workers, jobs, models, on-chain registration, and per-model network analytics. Independent, community-built (not an official LightChain package).",
   "type": "module",
   "main": "./dist/index.js",