Chrome Built-In AI APIs Bring Language Detection, Translation and Writing Tools to Browser

·2 views

Google Chrome now offers a suite of built-in AI APIs that allow developers to run AI workloads directly in the browser without managing any external model infrastructure. Stable Chrome on supported desktop devices includes the Language Detector, Translator, and Summarizer APIs, which require no experimental flags to use. The Prompt API, powered by Gemini Nano, supports general-purpose inference, while the Writer, Rewriter, and Proofreader APIs remain experimental and may need developer or origin trials to access. These APIs cover use cases ranging from real-time translation and content summarization to grammar correction and structured data extraction. Developers are advised to use runtime feature detection rather than relying on Chrome version numbers, as the APIs continue to evolve.

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)

Qodo Emerges as a Top AI Code Review Tool for VS Code Users

A hands-on review published on DEV Community on July 9 highlights Qodo as a standout AI-powered code review tool for Visual Studio Code. The tool is designed to boost developer productivity by automating and improving the code review process. Writer Anthony Max evaluated Qodo's features in a practical, real-world context targeting web developers and programmers. The article, which takes approximately four minutes to read, received strong community engagement with 110 reactions and multiple comments.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

How developers can cap costs when AI agents query large data warehouses autonomously

Autonomous AI agents can answer complex business questions by querying databases and iterating through trial-and-error, but this loop can become extremely expensive on large platforms like BigQuery. Unlike human analysts who scope queries carefully, agents may run unoptimized scans across entire columns, with a single 100TB query potentially costing over $600 at standard pricing. A developer using Google's Antigravity SDK and the Data Agent Kit built a cost-safe agent with two guardrails: one that dry-runs every query before execution to block oversized scans, and another that pauses the agent for human approval once token spending crosses a set budget. The post highlights that common assumptions — such as LIMIT clauses reducing scan costs — are false in columnar databases like BigQuery, where charges are based on bytes read rather than rows returned. The guardrails aim to make autonomous data agents practical in production environments without exposing organizations to runaway infrastructure costs.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Multi-Tenancy Explained: How One App Serves Many Businesses Securely

Multi-tenant architecture allows a single software system to serve multiple independent customers, called tenants, from a shared infrastructure while keeping their data strictly isolated. The most common approach uses shared database tables where every row is tagged with a tenant_id column to identify which tenant owns each piece of data. Engineers must layer multiple isolation mechanisms — including middleware, API gateways, and encryption — to ensure no tenant can access another's data. Supporting principles such as load balancing, idempotency, and dependency injection work together to keep the system reliable and maintainable at scale. Compared to single-tenant systems, multi-tenancy reduces infrastructure costs and maintenance overhead, though it demands significantly more engineering discipline to implement correctly.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Human editors caught AI inventing quotes and fake maxims twice in a single day

A writer and editor working with an AI writing tool discovered two separate fabrications within the same day during pre-publication review. In the first case, the AI presented what appeared to be a well-known trading maxim, but investigation revealed the line was actually a label the AI had written for its own internal summary notes, not a saying in real-world circulation. The second incident involved a direct quote attributed to a named real person, but a full-text search of the original source showed the exact wording never appeared there — the AI had silently substituted its own paraphrase inside quotation marks. Both errors were corrected by rewriting the passages as indirect speech, removing any false claim of an authoritative source. The author explains the root cause as structural: AI systems do not preserve the distinction between self-generated labels and actual quotations, making fabrications most likely to slip in precisely where prose is flowing smoothly.

0 comments Read more at DEV Community