Apple Adds Official Safari MCP Server in Technology Preview 247 for AI Debugging

·1 views

Apple shipped an official Safari MCP server with Safari Technology Preview 247 in early July 2026, marking the first time a major browser vendor has natively integrated Model Context Protocol support for AI-driven debugging. The server is built on safaridriver and exposes 17 tools covering navigation, DOM inspection, element interaction, network capture, console logging, and screenshots. It runs entirely on the user's machine with no data sent to Apple, and each AI session launches in an isolated window with no access to personal browser data such as cookies, logins, or autofill. Before this release, all MCP browser automation tools relied on Chromium, forcing Mac developers who prefer Safari to run a second browser solely for AI agent tasks. The server is currently available only in Safari Technology Preview and not yet in the stable Safari release.

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)

How a Fractional CTO Framework Helps Businesses Find Real AI Automation ROI

Many companies invest in AI tools but fail to see measurable gains in operational metrics such as processing times or error rates, a gap attributed to the lack of a structured integration strategy. A Senior IT Consultant and Fractional CTO with over a decade of experience has developed a step-by-step audit framework to help businesses move beyond mere AI adoption toward strategic integration. The framework begins with precisely mapping existing workflows before any automation is attempted. High-value automation candidates share three traits: high task volume, linear and predictable progression, and low judgment complexity. Vendor invoice processing is cited as a practical example, where repetitive manual data entry from PDF attachments into accounting software can be effectively replaced by OCR and API-based automation.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Zig Language Integrates Package Management Directly Into Its Build System

Zig, a systems programming language, manages dependencies entirely through its native build system via a build.zig configuration file, eliminating the need for external package managers like Cargo or Go Modules. The system uses checksum-based verification to ensure deterministic, reproducible builds across Windows, Linux, and macOS without manual configuration. Dependencies are fetched automatically, cached locally, and statically linked, producing standalone executables with no runtime overhead. Incremental builds reportedly reduce compile times by 40–60% by recompiling only modules affected by dependency changes. However, the approach still faces challenges including limited tooling maturity compared to established managers like npm, a Zig-specific configuration learning curve, and gaps in legacy C/C++ project support.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

MarketNow Lets MCP Server Developers Sell AI Tools for USDC or Card

MarketNow, a marketplace built by AliceLabs LLC, allows developers to list and sell Model Context Protocol (MCP) servers to AI agents and human developers at one-time prices between $0.99 and $9.99. The platform currently indexes over 8,500 MCP skills and sees around 675 weekly installs, yet most developers offer their tools for free. Sellers retain 80% of each sale, with payments processed via Stripe for card users or x402 and USDC on the Base Ethereum Layer 2 network for crypto transactions. The crypto payment option is aimed at developers in regions where Stripe is unavailable, including India, China, Africa, and Latin America, requiring only a wallet address to receive funds. Each listed tool undergoes a Sentinel L2 security audit covering code scanning, sandbox testing, and dependency checks before going live.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Four Experiments Show 'Deterministic AI Agent' Claims Fail at the Semantic Layer

A software developer on DEV Community ran four controlled experiments to test the core mechanisms promoted in popular 'production-grade AI agent' articles, which claim deterministic constraints can reliably govern LLM-based agent loops. The three mechanisms tested — lexical-overlap thresholds, temperature-0 evaluators, and phase gates — each proved only formally deterministic, breaking down when applied to real semantic judgments. Lexical overlap alone produced a 50% hard misclassification rate on 30 labeled pairs, including cases where a delete instruction was treated as a continuation of a writing task. The developer also attempted an upgraded fix to address these failures, but that too did not hold up under measurement. While the broader direction of wrapping LLM uncertainty in structured constraints is acknowledged as sound, the article warns that treating unvalidated mechanisms as solved engineering is misleading and potentially incident-grade in production.

0 comments Read more at DEV Community