every ai dev release, the day it lands
find the tool
for the thing you
are building right now
latest releases, docs and repos for models, agents, inference stacks and clis — with the numbers that say whether anyone actually uses them.
one field, one button — and it is free.
latest releases
all tools →Apache Spark 4.3 is a major update to the open-source unified analytics engine, now available in Databricks Runtime 19. It introduces stateful processing in Real-Time Mode for sub-150ms latency, history-preserving SCD Type 2 support for Auto CDC, and Project Feather for optimized small-data performance.
Exact is a specialized project associated with the Expo ecosystem, featuring a dedicated JavaScript/TypeScript runtime known as Ibex. It is designed to support advanced application capabilities, including precise scheduling and background execution tasks.
The HTML-in-Canvas API allows developers to render live, interactive, and accessible DOM subtrees directly into 2D, WebGL, and WebGPU canvas contexts. It bridges the gap between high-performance pixel-level graphics and the rich semantic capabilities of the web platform.
Whisker is a Rust-first, cross-platform framework for building native iOS, Android, web, and desktop applications. It utilizes a fine-grained reactive runtime and Taffy layout engine to provide high-performance UI without a virtual DOM.
Vercel Tracing is an observability tool that allows developers to search, inspect, and analyze request traces and spans directly from the Vercel dashboard or CLI. It provides deep visibility into application performance, enabling teams to identify bottlenecks and errors with support for Kibana Query Language (KQL).
Factory Wargames is a security program that simulates nation-state level attacks on an organization's infrastructure to test its defenses. Selected participants receive a controlled simulation, shared findings, and assistance from Factory's Droids to remediate identified vulnerabilities at no cost.
Aspen is an open-source, Git-compatible server backend designed to handle high-frequency pushes from hundreds of AI agents simultaneously. It scales by decoupling branch updates and serving pre-stored packfiles from object storage, effectively removing the performance bottlenecks found in traditional Git servers.
Intelligent Terminal is an experimental application that integrates AI agents natively into the Windows terminal experience. The v0.3 update introduces a new sidebar for managing tabs and agent sessions, along with a headless mode that allows tasks to continue running in the background after closing the terminal window.
Devin Review 2.0 is an AI-powered code review tool designed to automatically identify bugs and vulnerabilities in pull requests. It provides guided tours of PRs to assist human reviewers and can be configured to automatically fix issues, significantly reducing the bottleneck in the software development lifecycle.
Reflex is an MVVM-based mobile app architecture designed to enable AI agents to build, test, and drive mobile applications headlessly. By decoupling business logic from the UI and exposing state and actions through a standardized JSON contract, it allows for rapid feedback loops and cross-platform synchronization without requiring emulators.
AlloyDB Express is a developer-focused PostgreSQL environment designed for rapid testing and prototyping, featuring sub-2-second provisioning and scale-to-zero compute. It integrates enterprise-grade AI capabilities, such as vector search and Gemini-powered SQL generation, while maintaining strict security guardrails for developers and AI coding agents.
Flare is a specialized local inference server designed for NVIDIA SM120/SM121 GPUs, optimized for long-context conversations, coding agents, and repeated tool use. It utilizes a two-part architecture where Python-based 'Ignition' prepares models and a high-performance Rust-based 'Helios' runtime serves them without Python dependencies.
Paradev is a desktop orchestration platform that allows developers to run multiple CLI-based coding agents in parallel within isolated Git worktrees. It provides a unified workspace for managing agent tasks, terminal sessions, code reviews, and browser previews without the need for manual branch switching.
Qwen-Audio-Agent is a realtime voice runtime that enables full-duplex voice interaction for AI agents by separating foreground conversation from background task execution. It allows users to maintain a continuous, natural dialogue while the agent independently processes complex tasks, tool calls, or file operations in the background.
Three-Live is an AI-first 3D and spatial computing runtime that features an on-device JIT compiler to translate Three.js code into native rendering engines. It enables high-performance AR and XR experiences across Android (Filament/Jetpack XR), iOS/visionOS (RealityKit), and Web (WebGL2/WebXR).
Vercel Security Review is an AI-powered security scanning tool integrated into the Vercel Agent platform. It automatically analyzes code changes on every GitHub push to identify vulnerabilities like authorization flaws and exposed secrets, offering one-click remediation through the Vercel dashboard.
Sandstorm is an open-source sandbox fleet manager designed to run thousands of AI agent sandboxes on minimal hardware. It optimizes resource usage by checkpointing idle containers to disk and waking them instantly upon request, supporting both gVisor and Firecracker backends.
CoworkOS is an open-source agent harness designed for long-horizon, human-in-the-loop task execution. It enables AI agents and humans to collaborate on complex projects by separating task strategy from execution management, supporting persistent workspaces, and allowing shared interaction with professional tools like CAD and Blender.
Solid Glass is a design system and theme for AI applications built on top of shadcn/ui. It provides a refined, glass-inspired aesthetic that balances the opacity of flat design with the clarity of glass, specifically tailored for agent workspaces and operator consoles.
Hermit is an open-source framework designed for continual learning and recursive self-improvement of AI agents through harness adapter evolution. It allows users to adapt existing agents to new tasks by evolving external adapters without modifying the underlying model or original harness.
what we wrote this week
all posts →Vercel Adds Spend Management Capabilities to CLI
Vercel has introduced Spend Management features in its CLI, allowing users to monitor, set, and manage usage budgets directly from the terminal.
AWS Launches Continuum: An Autonomous AI Security Agent Designed to Find, Prove, and Patch Code Vulnerabilities at Machine Speed
AWS has announced the general availability of AWS Continuum, an AI-driven service designed to identify and remediate security vulnerabilities in code.
Microsoft Introduces Microsoft.UI.Reactor: An Experimental Framework for Building WinUI 3 Desktop Apps
Microsoft has released Microsoft.UI.Reactor, an experimental C#-first framework for building WinUI 3 desktop applications using the MVU pattern.
give us a follow
every release also goes out on x, linkedin and reddit the moment it lands.