Changelog

We are renaming Browser Rendering to Browser Run. The name Browser Rendering never fully captured what the product does. Browser Run lets you run full browser sessions on Cloudflare's global network, drive them with code or AI, record and replay sessions, crawl pages for content, debug in real time, and let humans intervene when your agent needs help.

Along with the rename, we have increased limits for Workers Paid plans and redesigned the Browser Run dashboard.

We have 4x-ed concurrency limits for Workers Paid plan users:

• Concurrent browsers per account: 30 → 120 per account
• New browser instances: 30 per minute → 1 per second
• REST API rate limits: recently increased from 3 to 10 requests per second

Rate limits across the limits page are now expressed in per-second terms, matching how they are enforced. No action is needed to benefit from the higher limits.

The redesigned dashboard ↗ now shows every request in a single Runs tab, not just browser sessions but also quick actions like screenshots, PDFs, markdown, and crawls. Filter by endpoint, view target URLs, status, and duration, and expand any row for more detail.

We are also shipping several new features:

• Live View, Human in the Loop, and Session Recordings - See what your agent is doing in real time, let humans step in when automation hits a wall, and replay any session after it ends.
• WebMCP - Websites can expose structured tools for AI agents to discover and call directly, replacing slow screenshot-analyze-click loops.

For the full story, read our Agents Week blog Browser Run: Give your agents a browser ↗.

When browser automation fails or behaves unexpectedly, it can be hard to understand what happened. We are shipping three new features in Browser Run (formerly Browser Rendering) to help:

• Live View for real-time visibility
• Human in the Loop for human intervention
• Session Recordings for replaying sessions after they end

Live View lets you see what your agent is doing in real time. The page, DOM, console, and network requests are all visible for any active browser session. Access Live View from the Cloudflare dashboard, via the hosted UI at live.browser.run, or using native Chrome DevTools.

When your agent hits a snag like a login page or unexpected edge case, it can hand off to a human instead of failing. With Human in the Loop, a human steps into the live browser session through Live View, resolves the issue, and hands control back to the script.

Today, you can step in by opening the Live View URL for any active session. Next, we are adding a handoff flow where the agent can signal that it needs help, notify a human to step in, then hand control back to the agent once the issue is resolved.

Session Recordings records DOM state so you can replay any session after it ends. Enable recordings by passing recording: true when launching a browser. After the session closes, view the recording in the Cloudflare dashboard under Browser Run > Runs, or retrieve via API using the session ID. Next, we are adding the ability to inspect DOM state and console output at any point during the recording.

To get started, refer to the documentation for Live View, Human in the Loop, and Session Recording.

Browser Run (formerly Browser Rendering) now supports WebMCP ↗ (Web Model Context Protocol), a new browser API from the Google Chrome team.

The Internet was built for humans, so navigating as an AI agent today is unreliable. WebMCP lets websites expose structured tools for AI agents to discover and call directly. Instead of slow screenshot-analyze-click loops, agents can call website functions like searchFlights() or bookTicket() with typed parameters, making browser automation faster, more reliable, and less fragile.

With WebMCP, you can:

• Discover website tools - Use navigator.modelContextTesting.listTools() to see available actions on any WebMCP-enabled site
• Execute tools directly - Call navigator.modelContextTesting.executeTool() with typed parameters
• Handle human-in-the-loop interactions - Some tools pause for user confirmation before completing sensitive actions

WebMCP requires Chrome beta features. We have an experimental pool with browser instances running Chrome beta so you can test emerging browser features before they reach stable Chrome. To start a WebMCP session, add lab=true to your /devtools/browser request:

curl -X POST "https://api.cloudflare.com/client/v4/accounts/{account_id}/browser-rendering/devtools/browser?lab=true&keep_alive=300000" \
  -H "Authorization: Bearer {api_token}"

Combined with the recently launched CDP endpoint, AI agents can also use WebMCP. Connect an MCP client to Browser Run via CDP, and your agent can discover and call website tools directly. Here's the same hotel booking demo, this time driven by an AI agent through OpenCode:

For a step-by-step guide, refer to the WebMCP documentation.

We are excited to announce two major capability upgrades for Agent Lee, the AI co-pilot built directly into the Cloudflare dashboard. Agent Lee is designed to understand your specific account configuration, and with this release, it moves from a passive advisor to an active assistant that can help you manage your infrastructure and visualize your data through natural language.

Agent Lee can now perform changes on your behalf across your Cloudflare account. Whether you need to update DNS records, modify SSL/TLS settings, or configure Workers routes, you can simply ask.

To ensure security and accuracy, every write operation requires explicit user approval. Before any change is committed, Agent Lee will present a summary of the proposed action in plain language. No action is taken until you select Confirm, and this approval requirement is enforced at the infrastructure level to prevent unauthorized changes.

Example requests:

• "Add an A record for blog.example.com pointing to 192.0.2.10."
• "Enable Always Use HTTPS on my zone."
• "Set the SSL mode for example.com to Full (strict)."

Understanding your traffic and security trends is now as easy as asking a question. Agent Lee now features Generative UI, allowing it to render inline charts and structured data visualizations directly within the chat interface using your actual account telemetry.

Example requests:

• "Show me a chart of my traffic over the last 7 days."
• "What does my error rate look like for the past 24 hours?"
• "Graph my cache hit rate for example.com this week."

These features are currently available in Beta for all users on the Free plan. To get started, log in to the Cloudflare dashboard ↗ and select Ask AI in the upper right corner.

To learn more about how to interact with your account using AI, refer to the Agent Lee documentation.

Browser Rendering now supports wrangler browser commands, letting you create, manage, and view browser sessions directly from your terminal, streamlining your workflow. Since Wrangler handles authentication, you do not need to pass API tokens in your commands.

The following commands are available:

The create command spins up a browser instance on Cloudflare's network and returns a session URL. Once created, you can connect to the session using any CDP-compatible client like Puppeteer, Playwright, or MCP clients to automate browsing, scrape content, or debug remotely.

wrangler browser create

Use --keepAlive to set the session keep-alive duration (60-600 seconds):

wrangler browser create --keepAlive 300

The view command auto-selects when only one session exists, or prompts for selection when multiple sessions are available.

All commands support --json for structured output, and because these are CLI commands, you can incorporate them into scripts to automate session management.

For full usage details, refer to the Wrangler commands documentation.

Outbound Workers for Sandboxes and Containers now support zero-trust credential injection, TLS interception, allow/deny lists, and dynamic per-instance egress policies. These features give platforms running agentic workloads full control over what leaves the sandbox, without exposing secrets to untrusted workloads, like user-generated code or coding agents.

Because outbound handlers run in the Workers runtime, outside the sandbox, they can hold secrets the sandbox never sees. A sandboxed workload can make a plain request, and credentials are transparently attached before a request is forwarded upstream.

For instance, you could run an agent in a sandbox and ensure that any requests it makes to Github are authenticated. But it will never be able to access the credentials:

export class MySandbox extends Sandbox {}

MySandbox.outboundByHost = {
	"github.com": (request: Request, env: Env, ctx: OutboundHandlerContext) => {
		const requestWithAuth = new Request(request);
		requestWithAuth.headers.set("x-auth-token", env.SECRET);
		return fetch(requestWithAuth);
	},
};

You can easily inject unique credentials for different instances by using ctx.containerId:

MySandbox.outboundByHost = {
	"my-internal-vcs.dev": async (
		request: Request,
		env: Env,
		ctx: OutboundHandlerContext,
	) => {
		const authKey = await env.KEYS.get(ctx.containerId);

		const requestWithAuth = new Request(request);
		requestWithAuth.headers.set("x-auth-token", authKey);
		return fetch(requestWithAuth);
	},
};

No token is ever passed into the sandbox. You can rotate secrets in the Worker environment and every request will pick them up immediately.

Outbound Workers now intercept HTTPS traffic. A unique ephemeral certificate authority (CA) and private key are created for each sandbox instance. The CA is placed into the sandbox and trusted by default. The ephemeral private key never leaves the container runtime sidecar process and is never shared across instances.

With TLS interception active, outbound Workers can act as a transparent proxy for both HTTP and HTTPS traffic.

Easily filter outbound traffic with allowedHosts and deniedHosts. When allowedHosts is set, it becomes a deny-by-default allowlist. Both properties support glob patterns.

export class MySandbox extends Sandbox {
	allowedHosts = ["github.com", "npmjs.org"];
}

Define named outbound handlers then apply or remove them at runtime using setOutboundHandler() or setOutboundByHost(). This lets you change egress policy for a running sandbox without restarting it.

export class MySandbox extends Sandbox {}

MySandbox.outboundHandlers = {
	allowHosts: async (req: Request, env: Env, ctx: OutboundHandlerContext ) => {
		const url = new URL(req.url);
		if (ctx.params.allowedHostnames.includes(url.hostname)) {
			return fetch(req);
		}
		return new Response(null, { status: 403 });
	},

	noHttp: async () => {
		return new Response(null, { status: 403 });
	},
};

Apply handlers programmatically from your Worker:

const sandbox = getSandbox(env.Sandbox, userId);

// Open network for setup
await sandbox.setOutboundHandler("allowHosts", {
	allowedHostnames: ["github.com", "npmjs.org"],
});
await sandbox.exec("npm install");

// Lock down after setup
await sandbox.setOutboundHandler("noHttp");

Handlers accept params, so you can customize behavior per instance without defining separate handler functions.

Upgrade to @cloudflare/containers@0.3.0 or @cloudflare/sandbox@0.8.9 to use these features.

For more details, refer to Sandbox outbound traffic and Container outbound traffic.

Browser Rendering now exposes the Chrome DevTools Protocol (CDP), the low-level protocol that powers browser automation. The growing ecosystem of CDP-based agent tools, along with existing CDP automation scripts, can now use Browser Rendering directly.

Any CDP-compatible client, including Puppeteer and Playwright, can connect from any environment, whether that is Cloudflare Workers, your local machine, or a cloud environment. All you need is your Cloudflare API key.

For any existing CDP script, switching to Browser Rendering is a one-line change:

const puppeteer = require("puppeteer-core");

const browser = await puppeteer.connect({
	browserWSEndpoint: `wss://api.cloudflare.com/client/v4/accounts/${ACCOUNT_ID}/browser-rendering/devtools/browser?keep_alive=600000`,
	headers: { Authorization: `Bearer ${API_TOKEN}` },
});

const page = await browser.newPage();
await page.goto("https://example.com");
console.log(await page.title());
await browser.close();

Additionally, MCP clients like Claude Desktop, Claude Code, Cursor, and OpenCode can now use Browser Rendering as their remote browser via the chrome-devtools-mcp ↗ package.

Here is an example of how to configure Browser Rendering for Claude Desktop:

{
	"mcpServers": {
		"browser-rendering": {
			"command": "npx",
			"args": [
				"-y",
				"chrome-devtools-mcp@latest",
				"--wsEndpoint=wss://api.cloudflare.com/client/v4/accounts/<ACCOUNT_ID>/browser-rendering/devtools/browser?keep_alive=600000",
				"--wsHeaders={\"Authorization\":\"Bearer <API_TOKEN>\"}"
			]
		}
	}
}

To get started, refer to the CDP documentation.

AI Search now supports CSS content selectors for website data sources. You can now define which parts of a crawled page are extracted and indexed by specifying CSS selectors paired with URL glob patterns.

Content selectors solve the problem of indexing only relevant content while ignoring navigation, sidebars, footers, and other boilerplate. When a page URL matches a glob pattern, only elements matching the corresponding CSS selector are extracted and converted to Markdown for indexing.

Configure content selectors via the dashboard or API:

curl "https://api.cloudflare.com/client/v4/accounts/{account_id}/ai-search/instances" \
  -H "Authorization: Bearer {api_token}" \
  -H "Content-Type: application/json" \
  -d '{
    "id": "my-ai-search",
    "source": "https://example.com",
    "type": "web-crawler",
    "source_params": {
      "web_crawler": {
        "parse_options": {
          "content_selector": [
            {
              "path": "**/blog/**",
              "selector": "article .post-body"
            }
          ]
        }
      }
    }
  }'

Selectors are evaluated in order, and the first matching pattern wins. You can define up to 10 content selector entries per instance.

For configuration details and examples, refer to the content selectors documentation.

AI Search now supports four additional Workers AI models across text generation and embedding.

GLM-4.7-Flash is a lightweight model from Zhipu AI with a 131,072 token context window, suitable for long-document summarization and retrieval tasks. Qwen3-30B-A3B is a mixture-of-experts model from Alibaba that activates only 3 billion parameters per forward pass, keeping inference fast while maintaining strong response quality.

Qwen3-Embedding-0.6B supports up to 4,096 input tokens, making it a good fit for indexing longer text chunks. EmbeddingGemma-300M from Google produces 768-dimension vectors and is optimized for low-latency embedding workloads.

All four models are available without additional provider keys since they run on Workers AI. Select them when creating or updating an AI Search instance in the dashboard or through the API.

For the full list of supported models, refer to Supported models.

We are partnering with Google to bring @cf/google/gemma-4-26b-a4b-it to Workers AI. Gemma 4 26B A4B is a Mixture-of-Experts (MoE) model built from Gemini 3 research, with 26B total parameters and only 4B active per forward pass. By activating a small subset of parameters during inference, the model runs almost as fast as a 4B-parameter model while delivering the quality of a much larger one.

Gemma 4 is Google's most capable family of open models, designed to maximize intelligence-per-parameter.

• Mixture-of-Experts architecture with 8 active experts out of 128 total (plus 1 shared expert), delivering frontier-level performance at a fraction of the compute cost of dense models
• 256,000 token context window for retaining full conversation history, tool definitions, and long documents across extended sessions
• Built-in thinking mode that lets the model reason step-by-step before answering, improving accuracy on complex tasks
• Vision understanding for object detection, document and PDF parsing, screen and UI understanding, chart comprehension, OCR (including multilingual), and handwriting recognition, with support for variable aspect ratios and resolutions
• Function calling with native support for structured tool use, enabling agentic workflows and multi-step planning
• Multilingual with out-of-the-box support for 35+ languages, pre-trained on 140+ languages
• Coding for code generation, completion, and correction

Use Gemma 4 26B A4B through the Workers AI binding (env.AI.run()), the REST API at /run or /v1/chat/completions, or the OpenAI-compatible endpoint.

For more information, refer to the Gemma 4 26B A4B model page.

AI Gateway now supports automatic retries at the gateway level. When an upstream provider returns an error, your gateway retries the request based on the retry policy you configure, without requiring any client-side changes.

You can configure the retry count (up to 5 attempts), the delay between retries (from 100ms to 5 seconds), and the backoff strategy (Constant, Linear, or Exponential). These defaults apply to all requests through the gateway, and per-request headers can override them.

This is particularly useful when you do not control the client making the request and cannot implement retry logic on the caller side. For more complex failover scenarios — such as failing across different providers — use Dynamic Routing.

For more information, refer to Manage gateways.

AI Search supports a wrangler ai-search command namespace. Use it to manage instances from the command line.

The following commands are available:

The create command guides you through setup, choosing a name, source type (r2 or web), and data source. You can also pass all options as flags for non-interactive use:

wrangler ai-search create my-instance --type r2 --source my-bucket

Use wrangler ai-search search to query an instance directly from the CLI:

wrangler ai-search search my-instance --query "how do I configure caching?"

All commands support --json for structured output that scripts and AI agents can parse directly.

For full usage details, refer to the Wrangler commands documentation.

AI Crawl Control now supports extending the underlying WAF rule with custom modifications. Any changes you make directly in the WAF custom rules editor — such as adding path-based exceptions, extra user agents, or additional expression clauses — are preserved when you update crawler actions in AI Crawl Control.

If the WAF rule expression has been modified in a way AI Crawl Control cannot parse, a warning banner appears on the Crawlers page with a link to view the rule directly in WAF.

For more information, refer to WAF rule management.

The latest release of the Agents SDK ↗ exposes agent state as a readable property, prevents duplicate schedule rows across Durable Object restarts, brings full TypeScript inference to AgentClient, and migrates to Zod 4.

Both useAgent (React) and AgentClient (vanilla JS) now expose a state property that reflects the current agent state. Previously, reading state required manually tracking it through the onStateUpdate callback.

React (useAgent)

const agent = useAgent({
	agent: "game-agent",
	name: "room-123",
});

// Read state directly — no separate useState + onStateUpdate needed
return <div>Score: {agent.state?.score}</div>;

// Spread for partial updates
agent.setState({ ...agent.state, score: (agent.state?.score ?? 0) + 10 });

const agent = useAgent<GameAgent, GameState>({
	agent: "game-agent",
	name: "room-123",
});

// Read state directly — no separate useState + onStateUpdate needed
return <div>Score: {agent.state?.score}</div>;

// Spread for partial updates
agent.setState({ ...agent.state, score: (agent.state?.score ?? 0) + 10 });

agent.state is reactive — the component re-renders when state changes from either the server or a client-side setState() call.

Vanilla JS (AgentClient)

const client = new AgentClient({
	agent: "game-agent",
	name: "room-123",
	host: "your-worker.workers.dev",
});

client.setState({ score: 100 });
console.log(client.state); // { score: 100 }

const client = new AgentClient<GameAgent>({
	agent: "game-agent",
	name: "room-123",
	host: "your-worker.workers.dev",
});

client.setState({ score: 100 });
console.log(client.state); // { score: 100 }

State starts as undefined and is populated when the server sends the initial state on connect (from initialState) or when setState() is called. Use optional chaining (agent.state?.field) for safe access. The onStateUpdate callback continues to work as before — the new state property is additive.

schedule() now supports an idempotent option that deduplicates by (type, callback, payload), preventing duplicate rows from accumulating when called in places that run on every Durable Object restart such as onStart().

Cron schedules are idempotent by default. Calling schedule("0 * * * *", "tick") multiple times with the same callback, expression, and payload returns the existing schedule row instead of creating a new one. Pass { idempotent: false } to override.

Delayed and date-scheduled types support opt-in idempotency:

import { Agent } from "agents";

class MyAgent extends Agent {
	async onStart() {
		// Safe across restarts — only one row is created
		await this.schedule(60, "maintenance", undefined, { idempotent: true });
	}
}

import { Agent } from "agents";

class MyAgent extends Agent {
	async onStart() {
		// Safe across restarts — only one row is created
		await this.schedule(60, "maintenance", undefined, { idempotent: true });
	}
}

Two new warnings help catch common foot-guns:

• Calling schedule() inside onStart() without { idempotent: true } emits a console.warn with actionable guidance (once per callback; skipped for cron and when idempotent is set explicitly).
• If an alarm cycle processes 10 or more stale one-shot rows for the same callback, the SDK emits a console.warn and a schedule:duplicate_warning diagnostics channel event.

AgentClient now accepts an optional agent type parameter for full type inference on RPC calls, matching the typed experience already available with useAgent.

const client = new AgentClient({
	agent: "my-agent",
	host: window.location.host,
});

// Typed call — method name autocompletes, args and return type inferred
const value = await client.call("getValue");

// Typed stub — direct RPC-style proxy
await client.stub.getValue();
await client.stub.add(1, 2);

const client = new AgentClient<MyAgent>({
	agent: "my-agent",
	host: window.location.host,
});

// Typed call — method name autocompletes, args and return type inferred
const value = await client.call("getValue");

// Typed stub — direct RPC-style proxy
await client.stub.getValue();
await client.stub.add(1, 2);

State is automatically inferred from the agent type, so onStateUpdate is also typed:

const client = new AgentClient({
	agent: "my-agent",
	host: window.location.host,
	onStateUpdate: (state) => {
		// state is typed as MyAgent's state type
	},
});

const client = new AgentClient<MyAgent>({
	agent: "my-agent",
	host: window.location.host,
	onStateUpdate: (state) => {
		// state is typed as MyAgent's state type
	},
});

Existing untyped usage continues to work without changes. The RPC type utilities (AgentMethods, AgentStub, RPCMethods) are now exported from agents/client for advanced typing scenarios. agents, @cloudflare/ai-chat, and @cloudflare/codemode now require zod ^4.0.0. Zod v3 is no longer supported.

• Turn serialization — onChatMessage() and _reply() work is now queued so user requests, tool continuations, and saveMessages() never stream concurrently.
• Duplicate messages on stop — Clicking stop during an active stream no longer splits the assistant message into two entries.
• Duplicate messages after tool calls — Orphaned client IDs no longer leak into persistent storage.

keepAlive() now uses a lightweight in-memory ref count instead of schedule rows. Multiple concurrent callers share a single alarm cycle. The @experimental tag has been removed from both keepAlive() and keepAliveWhile().

A new entry point @cloudflare/codemode/tanstack-ai adds support for TanStack AI's ↗ chat() as an alternative to the Vercel AI SDK's streamText():

import {
	createCodeTool,
	tanstackTools,
} from "@cloudflare/codemode/tanstack-ai";
import { chat } from "@tanstack/ai";

const codeTool = createCodeTool({
	tools: [tanstackTools(myServerTools)],
	executor,
});

const stream = chat({ adapter, tools: [codeTool], messages });

import { createCodeTool, tanstackTools } from "@cloudflare/codemode/tanstack-ai";
import { chat } from "@tanstack/ai";

const codeTool = createCodeTool({
	tools: [tanstackTools(myServerTools)],
	executor,
});

const stream = chat({ adapter, tools: [codeTool], messages });

To update to the latest version:

npm i agents@latest @cloudflare/ai-chat@latest

AI Search now offers new REST API endpoints for search and chat that use an OpenAI compatible format. This means you can use the familiar messages array structure that works with existing OpenAI SDKs and tools. The messages array also lets you pass previous messages within a session, so the model can maintain context across multiple turns.

Here is an example request to the Chat Completions endpoint using the new messages array format:

curl https://api.cloudflare.com/client/v4/accounts/{ACCOUNT_ID}/ai-search/instances/{NAME}/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer {API_TOKEN}" \
  -d '{
    "messages": [
      {
        "role": "system",
        "content": "You are a helpful documentation assistant."
      },
      {
        "role": "user",
        "content": "How do I get started?"
      }
    ]
  }'

For more details, refer to the AI Search REST API guide.

If you are using the previous AutoRAG API endpoints (/autorag/rags/), we recommend migrating to the new endpoints. The previous AutoRAG API endpoints will continue to be fully supported.

Refer to the migration guide for step-by-step instructions.

AI Search now supports public endpoints, UI snippets, and MCP, making it easy to add search to your website or connect AI agents.

Public endpoints allow you to expose AI Search capabilities without requiring API authentication. To enable public endpoints:

1. Go to AI Search in the Cloudflare dashboard. Go to AI Search ↗
2. Select your instance, and turn on Public Endpoint in Settings. For more details, refer to Public endpoint configuration.

UI snippets are pre-built search and chat components you can embed in your website. Visit search.ai.cloudflare.com ↗ to configure and preview components for your AI Search instance.

To add a search modal to your page:

<script
	type="module"
	src="https://<PUBLIC_ENDPOINT_ID>.search.ai.cloudflare.com/assets/v0.0.25/search-snippet.es.js"
></script>

<search-modal-snippet
	api-url="https://<PUBLIC_ENDPOINT_ID>.search.ai.cloudflare.com/"
	placeholder="Search..."
>
</search-modal-snippet>

For more details, refer to the UI snippets documentation.

The MCP endpoint allows AI agents to search your content via the Model Context Protocol. Connect your MCP client to:

https://<PUBLIC_ENDPOINT_ID>.search.ai.cloudflare.com/mcp

For more details, refer to the MCP documentation.

AI Search now supports custom metadata filtering, allowing you to define your own metadata fields and filter search results based on attributes like category, version, or any custom field you define.

You can define up to 5 custom metadata fields per AI Search instance. Each field has a name and data type (text, number, or boolean):

curl -X POST https://api.cloudflare.com/client/v4/accounts/{ACCOUNT_ID}/ai-search/instances \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer {API_TOKEN}" \
  -d '{
    "id": "my-instance",
    "type": "r2",
    "source": "my-bucket",
    "custom_metadata": [
      { "field_name": "category", "data_type": "text" },
      { "field_name": "version", "data_type": "number" },
      { "field_name": "is_public", "data_type": "boolean" }
    ]
  }'

How you attach metadata depends on your data source:

• R2 bucket: Set metadata using S3-compatible custom headers (x-amz-meta-*) when uploading objects. Refer to R2 custom metadata for examples.
• Website: Add <meta> tags to your HTML pages. Refer to Website custom metadata for details.

Use custom metadata fields in your search queries alongside built-in attributes like folder and timestamp:

curl https://api.cloudflare.com/client/v4/accounts/{ACCOUNT_ID}/ai-search/instances/{NAME}/search \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer {API_TOKEN}" \
  -d '{
    "messages": [
      {
        "content": "How do I configure authentication?",
        "role": "user"
      }
    ],
    "ai_search_options": {
      "retrieval": {
        "filters": {
          "category": "documentation",
          "version": { "$gte": 2.0 }
        }
      }
    }
  }'

Learn more in the metadata filtering documentation.

Workers AI is officially in the big models game. @cf/moonshotai/kimi-k2.5 is the first frontier-scale open-source model on our AI inference platform — a large model with a full 256k context window, multi-turn tool calling, vision inputs, and structured outputs. By bringing a frontier-scale model directly onto the Cloudflare Developer Platform, you can now run the entire agent lifecycle on a single, unified platform.

The model has proven to be a fast, efficient alternative to larger proprietary models without sacrificing quality. As AI adoption increases, the volume of inference is skyrocketing — now you can access frontier intelligence at a fraction of the cost.

• 256,000 token context window for retaining full conversation history, tool definitions, and entire codebases across long-running agent sessions
• Multi-turn tool calling for building agents that invoke tools across multiple conversation turns
• Vision inputs for processing images alongside text
• Structured outputs with JSON mode and JSON Schema support for reliable downstream parsing
• Function calling for integrating external tools and APIs into agent workflows

When an agent sends a new prompt, it resends all previous prompts, tools, and context from the session. The delta between consecutive requests is usually just a few new lines of input. Prefix caching avoids reprocessing the shared context, saving time and compute from the prefill stage. This means faster Time to First Token (TTFT) and higher Tokens Per Second (TPS) throughput.

Workers AI has done prefix caching, but we are now surfacing cached tokens as a usage metric and offering a discount on cached tokens compared to input tokens (pricing is listed on the model page).

curl -X POST \
  "https://api.cloudflare.com/client/v4/accounts/{account_id}/ai/run/@cf/moonshotai/kimi-k2.5" \
  -H "Authorization: Bearer {api_token}" \
  -H "Content-Type: application/json" \
  -H "x-session-affinity: ses_12345678" \
  -d '{
    "messages": [
      {
        "role": "system",
        "content": "You are a helpful assistant."
      },
      {
        "role": "user",
        "content": "What is prefix caching and why does it matter?"
      }
    ],
    "max_tokens": 2400,
    "stream": true
  }'

Some clients like OpenCode ↗ implement session affinity automatically. The Agents SDK ↗ starter also sets up the wiring for you.

For volumes of requests that exceed synchronous rate limits, you can submit batches of inferences to be completed asynchronously. We have revamped the Asynchronous Batch API with a pull-based system that processes queued requests as soon as capacity is available. With internal testing, async requests usually execute within 5 minutes, but this depends on live traffic.

The async API is the best way to avoid capacity errors in durable workflows. It is ideal for use cases that are not real-time, such as code scanning agents or research agents.

To use the asynchronous API, pass queueRequest: true:

// 1. Push a batch of requests into the queue
const res = await env.AI.run(
	"@cf/moonshotai/kimi-k2.5",
	{
		requests: [
			{
				messages: [{ role: "user", content: "Tell me a joke" }],
			},
			{
				messages: [{ role: "user", content: "Explain the Pythagoras theorem" }],
			},
		],
	},
	{ queueRequest: true },
);

// 2. Grab the request ID
const requestId = res.request_id;

// 3. Poll for the result
const result = await env.AI.run("@cf/moonshotai/kimi-k2.5", {
	request_id: requestId,
});

if (result.status === "queued" || result.status === "running") {
	// Retry by polling again
} else {
	return Response.json(result);
}

You can also set up event notifications to know when inference is complete instead of polling.

Use Kimi K2.5 through the Workers AI binding (env.AI.run()), the REST API at /run or /v1/chat/completions, AI Gateway, or via the OpenAI-compatible endpoint.

For more information, refer to the Kimi K2.5 model page, pricing, and prompt caching.

The latest releases of @cloudflare/codemode ↗ add a new MCP barrel export, remove ai and zod as required peer dependencies from the main entry point, and give you more control over the sandbox.

A new @cloudflare/codemode/mcp entry point provides two functions that wrap MCP servers with Code Mode:

• codeMcpServer({ server, executor }) — wraps an existing MCP server with a single code tool where each upstream tool becomes a typed codemode.* method.
• openApiMcpServer({ spec, executor, request }) — creates search and execute MCP tools from an OpenAPI spec with host-side request proxying and automatic $ref resolution.

import { codeMcpServer } from "@cloudflare/codemode/mcp";
import { DynamicWorkerExecutor } from "@cloudflare/codemode";

const executor = new DynamicWorkerExecutor({ loader: env.LOADER });

// Wrap an existing MCP server — all its tools become
// typed methods the LLM can call from generated code
const server = await codeMcpServer({ server: upstreamMcp, executor });

import { codeMcpServer } from "@cloudflare/codemode/mcp";
import { DynamicWorkerExecutor } from "@cloudflare/codemode";

const executor = new DynamicWorkerExecutor({ loader: env.LOADER });

// Wrap an existing MCP server — all its tools become
// typed methods the LLM can call from generated code
const server = await codeMcpServer({ server: upstreamMcp, executor });

Breaking change in v0.2.0: generateTypes and the ToolDescriptor / ToolDescriptors types have moved to @cloudflare/codemode/ai:

// Before
import { generateTypes } from "@cloudflare/codemode";

// After
import { generateTypes } from "@cloudflare/codemode/ai";

// Before
import { generateTypes } from "@cloudflare/codemode";

// After
import { generateTypes } from "@cloudflare/codemode/ai";

The main entry point (@cloudflare/codemode) no longer requires the ai or zod peer dependencies. It now exports:

The ai and zod peer dependencies are now optional — only required when importing from @cloudflare/codemode/ai.

DynamicWorkerExecutor now accepts an optional modules option to inject custom ES modules into the sandbox:

const executor = new DynamicWorkerExecutor({
	loader: env.LOADER,
	modules: {
		"utils.js": `export function add(a, b) { return a + b; }`,
	},
});

// Sandbox code can then: import { add } from "utils.js"

const executor = new DynamicWorkerExecutor({
	loader: env.LOADER,
	modules: {
		"utils.js": `export function add(a, b) { return a + b; }`,
	},
});

// Sandbox code can then: import { add } from "utils.js"

DynamicWorkerExecutor now normalizes code and sanitizes tool names internally. You no longer need to call normalizeCode() or sanitizeToolName() before passing code and functions to execute().

npm i @cloudflare/codemode@latest

See the Code Mode documentation for the full API reference.

AI Gateway now supports the cf-aig-collect-log-payload header, which controls whether request and response bodies are stored in logs. By default, this header is set to true and payloads are stored alongside metadata. Set this header to false to skip payload storage while still logging metadata such as token counts, model, provider, status code, cost, and duration.

This is useful when you need usage metrics but do not want to persist sensitive prompt or response data.

curl https://gateway.ai.cloudflare.com/v1/$ACCOUNT_ID/$GATEWAY_ID/openai/chat/completions \
  --header "Authorization: Bearer $TOKEN" \
  --header 'Content-Type: application/json' \
  --header 'cf-aig-collect-log-payload: false' \
  --data '{
    "model": "gpt-4o-mini",
    "messages": [
      {
        "role": "user",
        "content": "What is the email address and phone number of user123?"
      }
    ]
  }'

For more information, refer to Logging.

You can now set topK up to 50 when a Vectorize query returns values or full metadata. This raises the previous limit of 20 for queries that use returnValues: true or returnMetadata: "all".

Use the higher limit when you need more matches in a single query response without dropping values or metadata. Refer to the Vectorize API reference for query options and current topK limits.

We're excited to partner with NVIDIA to bring @cf/nvidia/nemotron-3-120b-a12b to Workers AI. NVIDIA Nemotron 3 Super is a Mixture-of-Experts (MoE) model with a hybrid Mamba-transformer architecture, 120B total parameters, and 12B active parameters per forward pass.

The model is optimized for running many collaborating agents per application. It delivers high accuracy for reasoning, tool calling, and instruction following across complex multi-step tasks.

Key capabilities:

• Hybrid Mamba-transformer architecture delivers over 50% higher token generation throughput compared to leading open models, reducing latency for real-world applications
• Tool calling support for building AI agents that invoke tools across multiple conversation turns
• Multi-Token Prediction (MTP) accelerates long-form text generation by predicting several future tokens simultaneously in a single forward pass
• 32,000 token context window for retaining conversation history and plan states across multi-step agent workflows

Use Nemotron 3 Super through the Workers AI binding (env.AI.run()), the REST API at /run or /v1/chat/completions, or the OpenAI-compatible endpoint.

For more information, refer to the Nemotron 3 Super model page.

Edit: this post has been edited to clarify crawling behavior with respect to site guidance.

You can now crawl an entire website with a single API call using Browser Rendering's new /crawl endpoint, available in open beta. Submit a starting URL, and pages are automatically discovered, rendered in a headless browser, and returned in multiple formats, including HTML, Markdown, and structured JSON. The endpoint is a verified bot (intermediary agent) that respects robots.txt and AI Crawl Control ↗ by default, making it easy for developers to comply with website rules, and making it less likely for crawlers to ignore web-owner guidance. This is great for training models, building RAG pipelines, and researching or monitoring content across a site.

Crawl jobs run asynchronously. You submit a URL, receive a job ID, and check back for results as pages are processed.

# Initiate a crawl
curl -X POST 'https://api.cloudflare.com/client/v4/accounts/{account_id}/browser-rendering/crawl' \
  -H 'Authorization: Bearer <apiToken>' \
  -H 'Content-Type: application/json' \
  -d '{
    "url": "https://blog.cloudflare.com/"
  }'

# Check results
curl -X GET 'https://api.cloudflare.com/client/v4/accounts/{account_id}/browser-rendering/crawl/{job_id}' \
  -H 'Authorization: Bearer <apiToken>'

Key features:

• Multiple output formats - Return crawled content as HTML, Markdown, and structured JSON (powered by Workers AI)
• Crawl scope controls - Configure crawl depth, page limits, and wildcard patterns to include or exclude specific URL paths
• Automatic page discovery - Discovers URLs from sitemaps, page links, or both
• Incremental crawling - Use modifiedSince and maxAge to skip pages that haven't changed or were recently fetched, saving time and cost on repeated crawls
• Static mode - Set render: false to fetch static HTML without spinning up a browser, for faster crawling of static sites
• Well-behaved bot - Honors robots.txt directives, including crawl-delay

Available on both the Workers Free and Paid plans.

Note: the /crawl endpoint cannot bypass Cloudflare bot detection or captchas, and self-identifies as a bot.

To get started, refer to the crawl endpoint documentation. If you are setting up your own site to be crawled, review the robots.txt and sitemaps best practices.

Real-time transcription in RealtimeKit now supports 10 languages with regional variants, powered by Deepgram Nova-3 running on Workers AI.

During a meeting, participant audio is routed through AI Gateway to Nova-3 on Workers AI — so transcription runs on Cloudflare's network end-to-end, reducing latency compared to routing through external speech-to-text services.

Set the language when creating a meeting via ai_config.transcription.language:

{
	"ai_config": {
		"transcription": {
			"language": "fr"
		}
	}
}

Supported languages include English, Spanish, French, German, Hindi, Russian, Portuguese, Japanese, Italian, and Dutch — with regional variants like en-AU, en-GB, en-IN, en-NZ, es-419, fr-CA, de-CH, pt-BR, and pt-PT. Use multi for automatic multilingual detection.

If you are building voice agents or real-time translation workflows, your agent can now transcribe in the caller's language natively — no extra services or routing logic needed.

• Transcription docs
• Nova-3 model page
• Workers AI
• AI Gateway

Browser Rendering REST API rate limits for Workers Paid plans have been increased from 3 requests per second (180/min) to 10 requests per second (600/min). No action is needed to benefit from the higher limit.

The REST API lets you perform common browser tasks with a single API call, and you can now do it at a higher rate.

• /content - Fetch HTML
• /screenshot - Capture screenshot
• /pdf - Render PDF
• /markdown - Extract Markdown from a webpage
• /snapshot - Take a webpage snapshot
• /scrape - Scrape HTML elements
• /json - Capture structured data using AI
• /links - Retrieve links from a webpage

If you use the Browser Sessions method, increases to concurrent browser and new browser limits are coming soon. Stay tuned.

For full details, refer to the Browser Rendering limits page.