Transcription
Hey devs and creators, this week we're showcasing top trending and best dev tool projects. This week, a mix of back-end engines, UI kits, storage tools, and coding helpers that can speed up your next project. You'll check out tools like Rust FS for S3 compatible storage, Fresh for Blazing Fast, Edge First We Apps, and River for clean AI agent streaming. Let's get started.
Welcome back to Manu AGI tutorials. Here we explore the exciting world of AI, latest AI tools for you. So don't forget to hit that subscribe button and the notification bell so you don't miss out on the latest AI insights. So let's start today's video.
Project number one, River safe, flexible streaming for AI agents. Imagine building an AI assistant that talks, reasons, and streams outputs. And you want that stream to behave reliably even if it pauses, disconnects, or needs type-safe handling. That's the promise of River, an open-source TypeScript library by a community developer that makes working with AI agent streaming responses much safer and more manageable. River gives you a clean, type-safe abstraction over agent streams. It handles the stream life cycle, lets you resume streams, and ensures that your code knows exactly what types of messages come in, avoiding messy callbacks or error-prone parsing. On top of that, River supports plugging into the kind of publish-subscribe or websocket-based real-time setup many AI agent frontends need, which means it can serve as a robust backbone when building real-time collaborative AI tools or agents. It targets developers and AI tool builders who want control, stability, and readability when handling streaming AI outputs. Perfect for chatbots, agent orchestration, or real-time AI-powered apps. If you want cleaner agent streams and less friction while coding, give River a try and feel how smoother your streaming code becomes.
Project number two, Rust FS. High-performance S3 compatible object storage. Imagine a storage backend that works like familiar cloud storage but runs on your own hardware. Fast and free of licensing headaches. That's what Rust FS offers. Rust FS is a free open-source distributed object storage system written in Rust. Rust FS implements a full S3 compatible API, so anything built to work with AWS S3 or S3 compatible clients can work unchanged. Its architecture is distributed and scalable. You can run it as a single node or a cluster and store objects, files, blobs in buckets across drives with peer-level nodes, no central metadata master, making it resilient and simple to deploy. What stands out is performance for small objects (4 kilobyte payloads). Rust FS benchmarks show 2.3x higher throughput than comparable systems like MinIO under similar conditions. Because Rust avoids a garbage-collected runtime, it avoids unpredictable pauses, a big advantage under heavy load or with many small files. Rust FS is licensed under a permissive Apache 2.0 license. Unlike MinIO's more restrictive AGPL, making it attractive for companies or projects that want to embed or modify storage code without license entanglement. It supports features such as object versioning, bucket lifecycle policies, encryption, cross-region/cross-site replication, and is suited for data lakes, AI workloads, big data storage, or any place needing robust object storage. Rust FS is built for developers, data engineers, companies running self-hosted infrastructure, or anyone who needs fast, scalable S3 compatible storage without vendor lock-in. It gives them speed, control, lower latency, and freedom. Try Rust FS and watch how storage feels lighter, faster, and more under your control.
Project number three, CAP, open-source screen recording and sharing. Imagine clicking record on your screen and instantly creating a polished video that you can share without juggling messy settings or worrying about cloud lock-in. That's what CAP does. It's a free open-source alternative to Loom, built to let you record, edit, and share screen recordings quickly with full control over your data. CAP's tech stack is modern and robust. The app relies on Rust for video capture and performance-sensitive parts, a React, Next.js plus TypeScript front-end, and uses a Tauri-based desktop wrapper, so it runs natively on Mac OS and Windows. No heavy Electron overhead. You can use it out of the box via its desktop client or optionally self-host the server-side with Docker or Railway. Store recordings on your own S3 or storage bucket and manage videos under your own domain. CAP supports high-quality 4K 60fps video, camera plus screen recording, editing tools, trim, layout control, and instant sharable links with privacy controls chosen by you. It's ideal for content creators, educators, teams, developers, or anyone who needs to record demos, tutorials, bug reports, or presentations while keeping ownership of content and avoiding vendor lock-in. Give CAP a try and see how effortlessly screen recording becomes part of your workflow.
Project number four, Animate UI, animated React UI component set. Imagine building a website and instead of manually coding every motion, you get ready-made components that already move, slide, fade, and reveal themselves naturally when used. That's what Animate UI offers. It's a free open-source library built with React, TypeScript, Tailwind CSS, Framer Motion, and the Shadcn UI CLI that provides a collection of animated UI primitives and components, making it easier to build fluid, modern interfaces without reinventing animation logic. With Animate UI, you get components like smoothly rotating or typing text, animated tab transitions, file list accordions, buttons with hover and tap motion, and more. You install it as you would with Shadcn UI, then import components such as a sliding number, rotating text container, animated tabs, or file explorer UI into your React project. Because it's built with modern web tech and distributed under MIT license, Animate UI works with standard React-based stacks and gives developers a fast, maintainable way to add polish and motion, making your UI feel alive, responsive, and pleasant to use. Try Animate UI and see how quickly your interface becomes smooth and expressive.
Project number five, Tanstack AI, a type-safe framework-agnostic AI SDK for developers. Imagine building an app that uses AI, text, images, audio, or even documents and doing that with one clean unified toolkit instead of juggling many different SDKs. That's what Tanstack AI brings. It is a free open-source SDK for JavaScript/TypeScript that enables developers to plug AI features into their applications simply and reliably. Tanstack AI supports multiple LLM providers such as OpenAI, Anthropic, Gemini, Alama via provider-agnostic adapters. It works in any JavaScript environment: React, Solid, VanillaJS. Thanks to a framework-agnostic core, one of its key strengths is type safety. You define tools, functions, APIs, etc. with schema definitions. The same definitions are used for both server-side and client-side code, reducing bugs and mismatches. Tanstack AI also supports streaming responses, multimodal content (images, audio, video, documents), and agent loop strategies, making it suitable for complex AI workflows and assistant-type apps. It is built for developers building web apps, AI-powered tools, chat interfaces, or any system needing a clean, maintainable bridge to AI models. With Tanstack AI, you get predictable APIs, flexibility to swap providers, and a unified foundation so AI becomes just another part of your code, not an afterthought. Give it a try and see how smoothly AI features integrate into your next application.
Project number six, Fresh, lightweight edge-first web framework for Deno. Imagine opening a browser and instantly seeing a fully rendered web page with no bulky JavaScript bundles, no build step, and blazing fast load times. That's what Fresh does. It's an open-source web framework built for the Deno runtime designed to deliver server-rendered pages efficiently while letting you add interactivity only where needed. Fresh uses a clever islands architecture, meaning it renders pages on the server by default and sends plain HTML to clients. You get full server-side rendering without shipping JavaScript unless a part of the page really needs it. If you do need interactivity, Fresh lets you embed small interactive components (islands) using Preact. Fresh offers built-in TypeScript support, file-system routing so root paths map to files like in some popular frameworks, and requires hardly any configuration. The framework skips a traditional build step. Your code runs as-is on Deno, which simplifies iteration and reduces deployment overhead. It suits developers who care about performance, simplicity, and small payloads. Especially useful for content sites, blogs, landing pages, or apps where fast load times and low bandwidth matter. Try Fresh and experience how smooth web delivery becomes, often with just HTML and a pinch of JS where needed.
Project number seven, Fizzy, a fresh open-source Kanban tool for small projects. Imagine you had a digital whiteboard where you could drop ideas, bugs, or tasks as colorful cards and manage them without endless menus or clutter. That's what Fizzy brings to life. Fizzy is a Kanban-style web-based project tracking tool created by Basecamp/37Signals offered under an open-source license, so you can self-host it or use it via their hosted service. At its core, Fizzy organizes work into boards made of simple columns and cards. Ideal for bugs, ideas, or small to mid-scale projects. It emphasizes simplicity, speed, and clarity over enterprise-style complexity. Cards can hold details, attachments, comments, and metadata like assignments, due dates, and you move them across columns as work progresses. Fizzy is built with modern web frameworks, a Ruby on Rails backend, a clean JS front-end, and internally uses efficient, minimal overhead design to stay responsive. It's meant for teams, creators, small businesses, or solo developers who want a light but expressive way to manage tasks without the overhead or corporate bloat of many popular tools. If you've ever felt overwhelmed by heavyweight project managers, give Fizzy a try and see how refreshing Kanban can be when you strip away the noise.
Thanks for watching. If you found these tools interesting, please like the video, subscribe to the channel, comment which repo you'd try first, and hit the bell to get notified about the next video. Don't forget to star the repos and try the demos. And if you know other great dev tools worth showcasing, drop their links in the comments for future videos.