All news with #cloudflare workers tag

Architecture of Remote Bindings for Local Worker Development

🚀 Cloudflare has made remote bindings generally available, letting local Workers connect to live resources such as R2 buckets, D1 and KV namespaces without deploying. Developers can enable a binding with "remote: true" in Wrangler v4.37.0 and use existing Wrangler OAuth credentials to access production data. The local workerd runtime proxies JS API calls to remote service bindings (including JSRPC via Cap’n Web websockets), and tooling like the Vite plugin and vitest-pool-workers can use utilities such as startRemoteProxySession to join remote sessions.

Cloudflare Introduces Isolated Testing for Workflows

🧪 Cloudflare has added local, isolated testing APIs for Workflows, enabling developers to introspect and mock workflow instances using the new cloudflare:test module. Available with @cloudflare/vitest-pool-workers v0.9.0+, the APIs (introspectWorkflowInstance and introspectWorkflow) let tests run offline inside the Workers runtime, mock step results and events, and preserve isolated storage for reliable, deterministic tests. This improves debug visibility, reduces flaky tests, and lets teams assert on intermediate steps without hitting external systems.

Cloudflare Workers: Automatic tracing now in open beta

🔍 Cloudflare announces an Open Beta for Workers tracing that provides automatic, out-of-the-box instrumentation with no code changes. Traces are visible in the Workers Observability dashboard alongside logs, and spans include timing, attributes, and error context. You can export OTLP-formatted traces and correlated logs to third-party providers like Honeycomb or Grafana. Enable tracing via wrangler.jsonc or the Cloudflare dashboard and join the beta to provide feedback.

Payload CMS on Cloudflare Workers with D1 and R2 Support

🧩 Deploy Payload to Cloudflare Workers in one click with a template that provisions D1 and R2 bindings. The Payload team ported the project to OpenNext and implemented custom adapters: a Drizzle-based adapter that maps D1 results for SQLite compatibility and an R2 storage adapter that uses bindings to avoid token management. They used Wrangler remote bindings for migrations and applied Hyperdrive and D1 read replicas to cut latency and improve global read performance.

Code Mode: Using MCP with Generated TypeScript APIs

🧩 Cloudflare introduces Code Mode, a new approach that converts Model Context Protocol (MCP) tool schemas into a generated TypeScript API so LLMs write code instead of emitting synthetic tool-call tokens. This lets models leverage broad exposure to real-world TypeScript, improving correctness when selecting and composing many or complex tools. Code Mode executes the generated code inside fast, sandboxed Cloudflare Workers isolates that expose only typed bindings to authorized MCP servers, preserving MCP's uniform authorization and discovery while reducing token overhead and orchestration latency.

Eliminating Cold Starts 2: Shard and Conquer Globally

🧊 Cloudflare describes a new Worker sharding technique that uses a consistent hash ring to route requests to existing Worker instances across a data center, reducing cold starts. The approach trades a sub-millisecond proxy hop for far fewer expensive cold starts, improving memory efficiency and latency. The system leverages Cap'n Proto RPC to implement optimistic forwarding, lazy capabilities, and seamless context transfer for nested Worker invocations.

Cloudflare Email Service Private Beta for Developers

📧 Cloudflare announced the private beta of its Email Sending capability, integrated into Workers so developers can send transactional emails directly from serverless code using a simple binding. The service complements existing Email Routing to provide a unified Email Service for both inbound and outbound flows, automates SPF/DKIM/DMARC setup to boost deliverability, and offers local testing, observability, and low-latency global delivery.

Cloudflare Workers Now Directly Connect to PlanetScale

🚀 Cloudflare Workers can now connect directly to PlanetScale Postgres and MySQL databases through a dashboard integration that links accounts and provisions an optimal Hyperdrive configuration. Built on Hyperdrive, the integration keeps connections warm, places pooled connections near your database, and can cache frequent read queries to reduce latency and database load. Credentials are managed securely, including a one-click password rotation, and the integration is accessible from both Cloudflare and PlanetScale dashboards to simplify full-stack app development.

R2 SQL Deep Dive: Serverless Queries over R2 Data Platform

⚡ R2 SQL is Cloudflare’s serverless query engine that runs SQL directly against Iceberg tables stored in R2, eliminating the need for Spark or Trino clusters. The Query Planner uses R2 Data Catalog metadata and multi-level stats to prune manifests, files, and Parquet row groups so only necessary bytes are read. Execution is distributed across Cloudflare’s network using Workers and query workers running Apache DataFusion, with results serialized via Apache Arrow. An ordered, streaming planning pipeline enables early termination for ORDER BY ... LIMIT queries; R2 SQL is currently available in open beta.

Cloudflare Workers: A Year of Node.js Compatibility

🔧 Over the past year Cloudflare has significantly expanded Node.js compatibility inside Workers, adding many core modules as native runtime features. The effort includes modules such as node:fs, node:crypto, node:http, node:net, and node:zlib, plus a virtual in-memory file system and native crypto via ncrypto. These changes reduce reliance on polyfills and tooling shims, improve performance and memory usage, and let popular npm packages and frameworks like Express run more seamlessly. Developers enable this with the nodejs_compat flag and can opt in or out of EOL-related APIs using granular compatibility flags.

Open-source VibeSDK for Self-hosted AI Coding Platforms

🚀 VibeSDK is an open-source platform that enables organizations to deploy a complete AI-powered "vibe coding" experience with one click, integrating LLMs, secure sandboxes, and scalable hosting. It provisions isolated development environments to safely execute AI-generated code, offers templates and live previews, and automates build, test, and deploy workflows. The SDK also provides multi-model routing, observability, and caching, plus one-click export to users' Cloudflare accounts or GitHub so teams retain control of code and costs.

Cloudflare Workers Launchpad Cohort #6 and Program Update

🚀 Cloudflare’s Workers Launchpad helps startups build and scale by providing cloud credits, technical mentorship, and direct access to product teams. Since expanding to a $2B funding commitment, the program has supported 145 startups across 23 countries, spanning AI, developer tools, and infrastructure. Cohort #6 introduces a new slate of founders and emphasizes hands-on collaboration, community, and pathways to funding and acquisition.

Make Websites Conversational with NLWeb and AutoRAG

🤖 Cloudflare offers a one-click path to conversational search by combining Microsoft’s NLWeb open standard with Cloudflare’s managed retrieval engine, AutoRAG. The integration crawls and indexes site content into R2 and a managed vector store, serves embeddings and inference via Workers AI, and exposes both a user-facing /ask endpoint and an agent-focused /mcp endpoint. Publishers get continuous re-indexing, controlled agent access, and observability through an AI Gateway, removing much of the infrastructure burden for conversational experiences.

Background Removal: Evaluating Image Segmentation Models

🧠 Cloudflare introduces background removal for Images, running a dichotomous image segmentation model on Workers AI to isolate subjects and produce soft saliency masks that map pixel opacity (0–255). The team evaluated U2-Net, IS-Net, BiRefNet, and SAM via the open-source rembg interface on the Humans and DIS5K datasets, prioritizing IoU and Dice metrics over pixel accuracy. BiRefNet-general achieved the best overall balance of fidelity and detail (IoU 0.87, Dice 0.92) while lightweight models were faster on modest GPUs and SAM was excluded for unprompted tasks. The feature is available in open beta through the Images API using the segment parameter and can be combined with other transforms or draw() overlays.

Cloudflare AI Gateway updates: unified billing, routing

🤖 Cloudflare’s AI Gateway refresh centralizes AI traffic management, offering unified billing, secure key storage, dynamic routing, and built-in security through a single endpoint. The update integrates Cloudflare Secrets Store for AES-encrypted BYO keys, provides an automatic normalization layer for requests/responses across providers, and introduces dashboard-driven Dynamic Routes for traffic splits, chaining, and limits. Native Firewall DLP scanning and configurable profiles add data protection controls, while partner access to 350+ models across six providers and a credits-based billing beta simplify procurement and cost management.

Cloudflare Workers AI Adds Leonardo and Deepgram Models

🚀 Cloudflare is expanding Workers AI to include closed-source partner models from Leonardo and Deepgram, bringing optimized image generation and real-time audio capabilities to the edge. The launch includes Leonardo's @cf/leonardo/phoenix-1.0 and @cf/leonardo/lucid-origin and Deepgram's @cf/deepgram/nova-3 and @cf/deepgram/aura-1. These models run on Cloudflare's low-latency GPU infrastructure and integrate with Workers, R2, Images, and Realtime for end-to-end developer workflows.

Cloudflare's Edge-Optimized LLM Inference Engine at Scale

⚡ Infire is Cloudflare’s new, Rust-based LLM inference engine built to run large models efficiently across a globally distributed, low-latency network. It replaces Python-based vLLM in scenarios where sandboxing and dynamic co-hosting caused high CPU overhead and reduced GPU utilization, using JIT-compiled CUDA kernels, paged KV caching, and fine-grained CUDA graphs to cut startup and runtime cost. Early benchmarks show up to 7% lower latency on H100 NVL hardware, substantially higher GPU utilization, and far lower CPU load while powering models such as Llama 3.1 8B in Workers AI.

How Cloudflare Runs More AI Models on Fewer GPUs with Omni

🤖 Cloudflare explains how Omni, an internal platform, consolidates many AI models onto fewer GPUs using lightweight process isolation, per-model Python virtual environments, and controlled GPU over-commitment. Omni’s scheduler spawns and manages model processes, isolates file systems with a FUSE-backed /proc/meminfo, and intercepts CUDA allocations to safely over-commit GPU RAM. The result is improved availability, lower latency, and reduced idle GPU waste.

Block Unsafe LLM Prompts with Firewall for AI at the Edge

🛡️ Cloudflare has integrated unsafe content moderation into Firewall for AI, using Llama Guard 3 to detect and block harmful prompts in real time at the network edge. The model-agnostic filter identifies categories including hate, violence, sexual content, criminal planning, and self-harm, and lets teams block or log flagged prompts without changing application code. Detection runs on Workers AI across Cloudflare's GPU fleet with a 2-second analysis cutoff, and logs record categories but not raw prompt text. The feature is available in beta to existing customers.

AI Prompt Protection: Contextual Control for GenAI Use

🔒 Cloudflare introduces AI prompt protection inside its Data Loss Prevention (DLP) product on Cloudflare One, designed to detect and secure data entered into web-based GenAI tools like Google Gemini, ChatGPT, Claude, and Perplexity. The capability captures both prompts and AI responses, classifies content and intent, and enforces identity-aware guardrails to enable safe, productive AI use without blanket blocking. Encrypted logging with customer-provided keys provides auditable records while preserving confidentiality.