Browsers Can Now Run AI Models Locally Using WebGPU, Cutting Cloud Costs

·4 views

As of 2026, developers can run large language model inference directly inside web browsers using WebGPU, eliminating the need for server-side API calls. Three main tools support this shift: WebLLM, built on Apache TVM and offering an OpenAI-compatible API; Transformers.js by Hugging Face, which supports vision, speech, and embeddings with a WASM fallback; and Chrome's native LanguageModel interface, which manages model weights at the browser level. WebGPU enables this by exposing GPU compute shaders to JavaScript, handling the parallel matrix operations that WebAssembly alone cannot efficiently perform. Chrome, Edge, and Safari already have robust WebGPU support, while Firefox still has partial gaps in certain features. Running models client-side keeps user data on-device, removes per-token cloud costs, and allows AI features to work offline.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Node 24 Runs TypeScript Natively by Stripping Types, Not Compiling Them

Node.js 24 can execute TypeScript files directly without a build step, using a bundled tool called Amaro that strips type annotations in place rather than compiling them. The feature, which became flag-free as of Node 23.6, works by replacing type annotations with blank spaces of equal length, preserving column positions and making error stack traces accurate without source maps. Because the mechanism is purely additive stripping, Node does not read tsconfig.json, perform type checking, or support TypeScript features that require code generation such as decorators. Developers can check which mode is active at runtime via process.features.typescript, and can disable stripping entirely using the --no-strip-types flag to verify compiled output is being used instead. The approach suits small services and scripts well but has documented limitations that can cause unexpected runtime errors if TypeScript-only features are relied upon.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Why Go Has Become a Go-To Language for High-Traffic Backend Systems

Go has gained popularity among companies handling millions of requests due to its practical advantages in building high-throughput backend services. Its goroutine-based concurrency model allows developers to manage thousands of simultaneous operations without the overhead of traditional OS threads. Go's compiled nature means applications run as native binaries, resulting in low memory usage, fast startup times, and simplified deployments with minimal dependencies. These traits make Go particularly cost-effective at scale, where infrastructure efficiency directly impacts operational expenses. While Go provides strong foundations, the article emphasizes that good architecture and engineering practices remain essential regardless of language choice.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

GitGuardian Adds S3 Bucket Scanning After IAM Key Breach Took 8 Minutes

AWS S3 buckets have become a significant security blind spot, accumulating years of logs, backups, and pipeline output that are rarely scanned for exposed credentials. A 2025 incident documented by Sysdig showed an attacker gaining administrative access to an AWS environment in just eight minutes after finding valid IAM keys in a public S3 bucket. AI tools are accelerating such attacks by automating reconnaissance and decision-making during intrusions. Security firm GitGuardian has responded by extending its secret scanning capabilities to cover AWS S3 buckets, including compressed archive formats, feeding findings into existing security workflows. Industry data suggests that 28% of cloud security incidents now originate outside code repositories, highlighting the need for scanning beyond traditional Git-based pipelines.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Developer ships Mac/iOS task app using raw SQLite C API, skipping Core Data and ORMs

A developer released KinetTask, a local-first task management app for Mac and iOS, built entirely on the SQLite3 C API in Swift 6.1 without using Core Data, SwiftData, or any ORM layer. The entire persistence layer spans roughly 600 lines of code, using WAL mode, prepared statements, and PRAGMA user_version for schema migrations. A significant bug was discovered post-launch where widget checkbox taps were silently lost because the main app and the widget were writing to two separate SQLite files in different sandboxed containers. The fix involved unifying both processes to a single App Group database file, enabling busy_timeout, and adding a snapshot-reconciliation pass to prevent stale in-memory cache from overwriting widget changes. The app also includes on-device AI features powered by Apple's foundation model or a GGUF model via llama.cpp, with an anti-filler gate added to prevent the model from generating unnecessary task structure.

0 comments Read more at DEV Community