fastmcp

10/ Want your own? FastMCP turns a Python function into an MCP server. Wrap a spreadsheet price lookup so your AI can answer "what do we charge for X?" Apache 2.0. Catch: you host it, and a paid API behind it still bills you. Free server, not free calls. jlowin/fastmcp

this is an alpha. skills support is experimental, and host support is still limited. already trying skills with fastmcp? check the release notes for api changes before upgrading. we’d love feedback from real use. github.com/PrefectHQ/fastmcp/releases/tag/v4.2.0a1

FastMCP 3.x/4.x: a transform tagging a tool "admin" hid it from tools/list, yet direct tools/call checked only the pre-transform policy and ran it for low-scope callers. Hidden isn't authorized. Upgrade to 3.4.10 or 4.1.2 and enforce auth in the handler. github.com/PrefectHQ/fastmcp/security/advisories/…

Sending an empty list to an MCP tool? FastMCP 4.1.4 fixes generated CLIs that dropped empty arrays. An explicit [] could otherwise silently use the server's nonempty default. Check scripts where [] and a missing argument mean different things. github.com/PrefectHQ/fastmcp/releases/tag/v4.1.4

FastMCP 4.0 to 4.1.1: its M2M OAuth client sent the client secret to whatever auth server the MCP server advertised, so a hostile MCP URL could collect it. 4.1.2 adds issuer=, but upgrading alone doesn't protect you. Set issuer, then rotate secrets. github.com/PrefectHQ/fastmcp/security/advisories/…

Heads up if your agents connect to MCP servers you don't control. FastMCP had a critical bug where a sketchy server could run code on your client with a crafted tool schema. Update to 3.4.8 or 4.0.11, there's no 2.x patch github.com/PrefectHQ/fastmcp/security/advisories/…

Build Model Context Protocol Servers in 5 Lines of Python with FastMCP! 💻⚡ (Open Source) Today we brought another repo called FastMCP — the high-level Python framework for Model Context Protocol! 💻⚡ Stop writing hundreds of lines of low-level JSON-RPC boilerplate just to give AI agents access to your local tools. In Day 39 of our GitHub Repo Series on AI Skill Vault, we explore how FastMCP automatically infers JSON schemas from standard Python type hints and docstrings to launch protocol servers in under 5 lines of code. ⭐ GitHub Repository: jlowin/fastmcp 💻 Language: Python 🛠️ Key Features: Zero-boilerplate tool definitions, automatic Pydantic schema generation, built-in image & resource streaming, and instant interactive debugging with `fastmcp dev`. 📌 TIMESTAMPS: 0:00 - Today we brought another repo called FastMCP 0:10 - How It Helps: Deploying MCP servers in under 5 lines of Python 0:20 - Code Demo: `pip install fastmcp`, `@mcp.tool()` decorators & running `fastmcp dev` 0:30 - Subscribe to AI Skill Vault Want more practical AI repos, agents and automation? Subscribe to AI Skill Vault. #github #developer #ai #python #fastmcp #mcp #modelcontextprotocol #aiagents #opensource #coding #aiskillvault

5 Free Open-Source AI Tools Actually Worth Trying Try Arcade, the runtime for MCP: arcade.dev.plug.dev Written guide & resources: lukejbyrne.com Some of the most interesting AI tools are projects you can download, run and change yourself. In this video I walk through five open-source projects: OpenDots for AI coworkers, FastMCP for custom AI tools, Impeccable for website design, REA for understanding how software works, and Chatterbox for speech. I cover what each one does, a useful example, and where to start. You'll see ideas for turning an editing workflow into an MCP tool, improving a landing page, and comparing two voice deliveries. Open source gives you more control, but models, hosting and compute can still cost money. What you'll learn: — Build an AI coworker workspace with OpenDots: CopilotKit/OpenDots — Turn Python functions into AI tools with FastMCP: PrefectHQ/fastmcp — Critique and polish website designs with Impeccable: pbakaus/impeccable — Inspect software and recreate a specific behaviour with REA: rea.tools — Compare speech delivery with Chatterbox: resemble-ai/chatterbox Watch next: Codex Masterclass: youtu.be Claude Code Masterclass: youtu.be Build Your First SaaS: youtu.be My Links: Create with AI: skool.com Typeless, the voice dictation I use every day: typeless.com Script & package your videos with AI: scripti.ai Raw tech news with no tracking: techl8.com Chapters: 0:00 Five open-source AI tools 0:29 OpenDots — AI coworkers 2:27 FastMCP — custom AI tools 4:35 Arcade — sponsor 5:38 Impeccable — better website design 7:23 REA — understand how software works 9:21 Chatterbox — open-source speech 9:58 Chatterbox voice comparison 11:32 Recap and which tool to try first This video is sponsored by Arcade. Contact: luke@byrnemedia.group [ar] 5 أدوات ذكاء اصطناعي مفتوحة المصدر ومجانية تستحق التجربة فعلاً جرّب Arcade، بيئة تشغيل MCP: arcade.dev.plug.dev بعضٌ من أكثر أدوات الذكاء الاصطناعي إثارةً للاهتمام هي مشاريع يمكنك تحميلها وتشغيلها وتعديلها بنفسك. في هذا الفيديو، أستعرض خمسة مشاريع مفتوحة المصدر: OpenDots لزملاء العمل الذين يستخدمون الذكاء الاصطناعي، وFastMCP لأدوات الذكاء الاصطناعي المخصصة، وImpeccable لتصميم مواقع الويب، وREA لفهم كيفية عمل البرامج، وChatterbox لمعالجة الكلام. سأشرح وظيفة كل مشروع، وأقدم مثالًا عمليًا، وأوضح كيفية البدء. ستتعرف على أفكار لتحويل سير عمل التحرير إلى أداة MCP، وتحسين صفحة هبوط، ومقارنة أداءين صوتيين. يمنحك المصدر المفتوح مزيدًا من التحكم، ولكن النماذج والاستضافة والحوسبة قد تكون مكلفة. ما ستتعلمه: — بناء بيئة عمل لزميل ذكاء اصطناعي باستخدام OpenDots: CopilotKit/OpenDots — تحويل دوال بايثون إلى أدوات ذكاء اصطناعي باستخدام FastMCP: PrefectHQ/fastmcp — تقييم وصقل تصميمات مواقع الويب باستخدام Impeccable: pbakaus/impeccable — فحص البرامج وإعادة إنشاء سلوك محدد باستخدام REA: rea.tools — مقارنة أداء الكلام باستخدام Chatterbox: resemble-ai/chatterbox شاهد أيضًا: دورة Codex التدريبية المتقدمة: youtu.be دورة Claude Code التدريبية المتقدمة: youtu.be بناء أول تطبيق SaaS لك: youtu.be روابطي: الإبداع باستخدام الذكاء الاصطناعي: skool.com Typeless، برنامج الإملاء الصوتي الذي أستخدمه يوميًا: typeless.com اكتب نصوص فيديوهاتك وقم بتجميعها باستخدام الذكاء الاصطناعي: scripti.ai أخبار تقنية مباشرة بدون تتبع: techl8.com الفصول: 0:00 خمس أدوات ذكاء اصطناعي مفتوحة المصدر 0:29 OpenDots - زملاء عمل من الذكاء الاصطناعي 2:27 FastMCP - أدوات ذكاء اصطناعي مخصصة 4:35 Arcade - الراعي 5:38 Impeccable - تصميم مواقع ويب أفضل 7:23 REA - فهم كيفية عمل البرامج 9:21 Chatterbox - برنامج كلام مفتوح المصدر 9:58 مقارنة صوتية لـ Chatterbox 11:32 ملخص وأفضل أداة للتجربة أولًا هذا الفيديو برعاية Arcade. للتواصل: luke@byrnemedia.group [bn] ৫টি ফ্রি ওপেন-সোর্স এআই টুল যা সত্যিই ব্যবহার করে দেখার মতো MCP-এর রানটাইম Arcade ব্যবহার করে দেখুন: arcade.dev.plug.dev সবচেয়ে আকর্ষণীয় কিছু AI টুল হলো এমন সব প্রজেক্ট যা আপনি নিজে ডাউনলোড, রান এবং পরিবর্তন করতে পারেন। এই ভিডিওতে আমি পাঁচটি ওপেন-সোর্স প্রজেক্ট নিয়ে আলোচনা করেছি: AI সহকর্মীর জন্য OpenDots, কাস্টম AI টুলের জন্য FastMCP, ওয়েবসাইট ডিজাইনের জন্য Impeccable, সফটওয়্যার কীভাবে কাজ করে তা বোঝার জন্য REA, এবং স্পিচের জন্য Chatterbox। আমি প্রতিটি কী কাজ করে, একটি দরকারি উদাহরণ এবং কোথা থেকে শুরু করতে হবে তা তুলে ধরেছি। আপনি একটি এডিটিং ওয়ার্কফ্লোকে MCP টুলে পরিণত করার, একটি ল্যান্ডিং পেজ উন্নত করার এবং দুটি ভয়েস ডেলিভারির তুলনা করার ধারণা দেখতে পাবেন। ওপেন সোর্স আপনাকে আরও বেশি নিয়ন্ত্রণ দেয়, কিন্তু মডেল, হোস্টিং এবং কম্পিউটের জন্য এখনও অর্থ খরচ হতে পারে। আপনি যা শিখবেন: OpenDots দিয়ে একটি AI সহকর্মী ওয়ার্কস্পেস তৈরি করুন: CopilotKit/OpenDots — FastMCP দিয়ে পাইথন ফাংশনগুলোকে AI টুলে পরিণত করুন: PrefectHQ/fastmcp — Impeccable দিয়ে ওয়েবসাইট ডিজাইন সমালোচনা ও পরিমার্জন করুন: pbakaus/impeccable — REA দিয়ে সফটওয়্যার পরিদর্শন করুন এবং একটি নির্দিষ্ট আচরণ পুনরায় তৈরি করুন: rea.tools — Chatterbox দিয়ে স্পিচ ডেলিভারি তুলনা করুন: resemble-ai/chatterbox এরপর দেখুন: Codex মাস্টারক্লাস: youtu.be Claude Code মাস্টারক্লাস: youtu.be আপনার প্রথম SaaS তৈরি করুন: youtu.be আমার লিঙ্কসমূহ: Create with AI: skool.com Typeless, the voice dictation I use every দিন: typeless.com এআই দিয়ে আপনার ভিডিওর স্ক্রিপ্ট ও প্যাকেজিং করুন: scripti.ai ট্র্যাকিং ছাড়া খাঁটি প্রযুক্তি সংবাদ: techl8.com অধ্যায়সমূহ: ০:০০ পাঁচটি ওপেন-সোর্স এআই টুল ০:২৯ ওপেনডটস — এআই সহকর্মী ২:২৭ ফাস্টএমসিপি — কাস্টম এআই টুল ৪:৩৫ আর্কেড — স্পনসর ৫:৩৮ ইম্পেকেবল — উন্নত ওয়েবসাইট ডিজাইন ৭:২৩ আরইএ — সফটওয়্যার কীভাবে কাজ করে তা বুঝুন ৯:২১ চ্যাটারবক্স — ওপেন-সোর্স স্পিচ ৯:৫৮ চ্যাটারবক্স ভয়েস তুলনা ১১:৩২ সারসংক্ষেপ এবং কোন টুলটি প্রথমে চেষ্টা করবেন এই ভিডিওটি আর্কেড দ্বারা স্পনসরকৃত। যোগাযোগ: luke@byrnemedia.group [de-DE] 5 kostenlose Open-Source-KI-Tools, die sich wirklich lohnen Probieren Sie Arcade, die Laufzeitumgebung für MCP, aus: arcade.dev.plug.dev Einige der interessantesten KI-Tools sind Projekte, die Sie herunterladen, ausführen und selbst anpassen können. In diesem Video stelle ich fünf Open-Source-Projekte vor: OpenDots für KI-gestütztes Arbeiten, FastMCP für benutzerdefinierte KI-Tools, Impeccable für Webdesign, REA zum Verständnis der Funktionsweise von Software und Chatterbox für Spracherkennung. Ich erkläre die Funktionen jedes Projekts, zeige ein praktisches Beispiel und gebe Tipps für den Einstieg. Sie erhalten Anregungen, wie Sie einen Bearbeitungs-Workflow in ein MCP-Tool umwandeln, eine Landingpage optimieren und zwei Sprachaufnahmen vergleichen können. Open Source bietet Ihnen mehr Kontrolle, aber Modelle, Hosting und Rechenleistung können dennoch Kosten verursachen. Was Sie lernen werden: — Erstellen Sie einen KI-gestützten Arbeitsbereich mit OpenDots: CopilotKit/OpenDots — Verwandeln Sie Python-Funktionen mit FastMCP in KI-Tools: PrefectHQ/fastmcp Bewerten und optimieren Sie Webdesigns mit Impeccable: pbakaus/impeccable — Untersuchen Sie Software und reproduzieren Sie bestimmtes Verhalten mit REA: rea.tools — Vergleichen Sie Sprachausgabe mit Chatterbox: resemble-ai/chatterbox Weitere Videos: Codex Masterclass: youtu.be Claude Code Masterclass: youtu.be Erstellen Sie Ihre erste SaaS-Lösung: youtu.be Meine Links: Erstellen mit KI: skool.com Typeless, die Stimme Diktierfunktion, die ich täglich nutze: typeless.com Skripte und verpacke deine Videos mit KI: scripti.ai Unverbindliche Tech-News ohne Tracking: techl8.com Kapitel: 0:00 Fünf Open-Source-KI-Tools 0:29 OpenDots – KI-Kollegen 2:27 FastMCP – Benutzerdefinierte KI-Tools 4:35 Arcade – Sponsor 5:38 Impeccable – Besseres Webdesign 7:23 REA – Software verstehen 9:21 Chatterbox – Open-Source-Spracherkennung 9:58 Chatterbox-Sprachvergleich 11:32 Zusammenfassung und welches Tool du zuerst ausprobieren solltest Dieses Video wird von Arcade gesponsert. Kontakt: luke@byrnemedia.group [es-US] 5 herramientas de IA de código abierto gratuitas que realmente vale la pena probar Prueba Arcade, el entorno de ejecución para MCP: arcade.dev.plug.dev Algunas de las herramientas de IA más interesantes son proyectos que puedes descargar, ejecutar y modificar tú mismo. En este video, explico cinco proyectos de código abierto: OpenDots para la colaboración con IA, FastMCP para herramientas de IA personalizadas, Impeccable para el diseño web, REA para comprender el funcionamiento del software y Chatterbox para el reconocimiento de voz. Explico qué hace cada uno, un ejemplo práctico y por dónde empezar. Verás ideas para convertir un flujo de trabajo de edición en una herramienta MCP, mejorar una página de destino y comparar dos grabaciones de voz. El código abierto te da más control, pero los modelos, el alojamiento y la computación aún pueden tener un costo. Lo que aprenderás: — Crea un espacio de trabajo colaborativo con IA usando OpenDots: CopilotKit/OpenDots — Convierte funciones de Python en herramientas de IA con FastMCP: PrefectHQ/fastmcp — Analiza y perfecciona diseños web con Impeccable: pbakaus/impeccable — Inspecciona software y recrea un comportamiento específico con REA: rea.tools — Compara la calidad de la voz con Chatterbox: resemble-ai/chatterbox Ver a continuación: Clase magistral de Codex: youtu.be Clase magistral de Claude Code: youtu.be Crea tu primer SaaS: youtu.be Mis enlaces: Crea con IA: skool.com Typeless, la voz Dictado que uso a diario: typeless.com Crea y edita tus vídeos con IA: scripti.ai Noticias tecnológicas sin seguimiento: techl8.com Capítulos: 0:00 Cinco herramientas de IA de código abierto 0:29 OpenDots: herramientas de IA para trabajar 2:27 FastMCP: herramientas de IA personalizadas 4:35 Arcade: patrocinador 5:38 Impeccable: mejor diseño web 7:23 REA: cómo funciona el software 9:21 Chatterbox: reconocimiento de voz de código abierto 9:58 Comparativa de voces de Chatterbox 11:32 Resumen y qué herramienta probar primero Este vídeo está patrocinado por Arcade. Contacto: luke@byrnemedia.group [fr-FR] 5 outils d'IA open-source gratuits qui valent vraiment le détour Essayez Arcade, l'environnement d'exécution pour MCP : arcade.dev.plug.dev Parmi les outils d'IA les plus intéressants, certains sont des projets que vous pouvez télécharger, exécuter et personnaliser. Dans cette vidéo, je vous présente cinq projets open source : OpenDots pour les assistants IA collaboratifs, FastMCP pour les outils d'IA personnalisés, Impeccable pour la conception de sites web, REA pour comprendre le fonctionnement des logiciels et Chatterbox pour la reconnaissance vocale. J'explique le fonctionnement de chacun, je donne un exemple concret et je vous indique par où commencer. Vous découvrirez des idées pour transformer un flux de travail de montage en un outil MCP, améliorer une page d'accueil et comparer deux performances vocales. L'open source vous offre plus de contrôle, mais les modèles, l'hébergement et la puissance de calcul peuvent tout de même engendrer des coûts. Ce que vous apprendrez : — Créer un espace de travail collaboratif IA avec OpenDots : CopilotKit/OpenDots — Transformer des fonctions Python en outils d’IA avec FastMCP : PrefectHQ/fastmcp — Analyser et optimiser des designs de sites web avec Impeccable : pbakaus/impeccable — Inspecter des logiciels et recréer un comportement spécifique avec REA : rea.tools — Comparer la restitution vocale avec Chatterbox : resemble-ai/chatterbox À voir ensuite : Masterclass Codex : youtu.be Masterclass Claude Code : youtu.be Créer votre premier SaaS : youtu.be Mes liens : Créer avec l’IA : skool.com Typeless, la voix Dictée que j'utilise au quotidien : typeless.com Créez vos scripts et vos montages vidéo avec l'IA : scripti.ai Actualités tech brutes et anonymes : techl8.com Chapitres : 0:00 Cinq outils d'IA open source 0:29 OpenDots — Collaborateurs IA 2:27 FastMCP — Outils d'IA personnalisés 4:35 Arcade — Sponsor 5:38 Impeccable — Un meilleur design de site web 7:23 REA — Comprendre le fonctionnement des logiciels 9:21 Chatterbox — Synthèse vocale open source 9:58 Comparaison des voix Chatterbox 11:32 Récapitulatif et quel outil essayer en premier Cette vidéo est sponsorisée par Arcade. Contact : luke@byrnemedia.group [hi] 5 मुफ्त ओपन-सोर्स AI टूल्स जिन्हें आज़माना वाकई फायदेमंद है MCP के लिए रनटाइम Arcade को आज़माएँ: arcade.dev.plug.dev कुछ सबसे दिलचस्प AI टूल ऐसे प्रोजेक्ट हैं जिन्हें आप डाउनलोड कर सकते हैं, चला सकते हैं और खुद बदल सकते हैं। इस वीडियो में मैं पाँच ओपन-सोर्स प्रोजेक्ट्स के बारे में बता रहा हूँ: AI सहकर्मियों के लिए OpenDots, कस्टम AI टूल के लिए FastMCP, वेबसाइट डिज़ाइन के लिए Impeccable, सॉफ़्टवेयर कैसे काम करता है यह समझने के लिए REA और स्पीच के लिए Chatterbox। मैं इनमें से प्रत्येक के कार्य, एक उपयोगी उदाहरण और शुरुआत करने के तरीके के बारे में बता रहा हूँ। आपको एडिटिंग वर्कफ़्लो को MCP टूल में बदलने, लैंडिंग पेज को बेहतर बनाने और दो अलग-अलग आवाज़ों की तुलना करने के तरीके भी देखने को मिलेंगे। ओपन सोर्स आपको ज़्यादा नियंत्रण देता है, लेकिन मॉडल, होस्टिंग और कंप्यूटिंग में अभी भी पैसे लग सकते हैं। आप सीखेंगे: — OpenDots के साथ एक AI सहकर्मी कार्यक्षेत्र बनाएं: CopilotKit/OpenDots — FastMCP के साथ Python फ़ंक्शंस को AI टूल्स में बदलें: PrefectHQ/fastmcp — Impeccable के साथ वेबसाइट डिज़ाइनों की समीक्षा और सुधार करें: pbakaus/impeccable — REA के साथ सॉफ़्टवेयर का निरीक्षण करें और एक विशिष्ट व्यवहार को पुनः उत्पन्न करें: rea.tools — Chatterbox के साथ भाषण वितरण की तुलना करें: resemble-ai/chatterbox अगला वीडियो देखें: Codex मास्टरक्लास: youtu.be Claude कोड मास्टरक्लास: youtu.be अपना पहला SaaS बनाएं: youtu.be मेरे लिंक: Create with AI: skool.com Typeless, वो वॉइस डिक्टेशन टूल जिसका मैं हर दिन इस्तेमाल करता हूँ: typeless.com AI की मदद से अपने वीडियो की स्क्रिप्ट और पैकेजिंग करें: scripti.ai बिना ट्रैकिंग के रॉ टेक न्यूज़: techl8.com अध्याय: 0:00 पाँच ओपन-सोर्स AI टूल 0:29 OpenDots — AI सहकर्मी 2:27 FastMCP — कस्टम AI टूल 4:35 Arcade — प्रायोजक 5:38 Impeccable — बेहतर वेबसाइट डिज़ाइन 7:23 REA — समझें कि सॉफ़्टवेयर कैसे काम करता है 9:21 Chatterbox — ओपन-सोर्स स्पीच 9:58 Chatterbox वॉइस तुलना 11:32 सारांश और सबसे पहले कौन सा टूल आज़माना चाहिए यह वीडियो Arcade द्वारा प्रायोजित है। संपर्क करें: luke@byrnemedia.group [id] 5 Alat AI Open-Source Gratis yang Layak Dicoba Coba Arcade, runtime untuk MCP: arcade.dev.plug.dev Beberapa alat AI yang paling menarik adalah proyek yang dapat Anda unduh, jalankan, dan ubah sendiri. Dalam video ini, saya akan membahas lima proyek open-source: OpenDots untuk rekan kerja AI, FastMCP untuk alat AI kustom, Impeccable untuk desain situs web, REA untuk memahami cara kerja perangkat lunak, dan Chatterbox untuk pengenalan suara. Saya akan membahas fungsi masing-masing, contoh yang bermanfaat, dan dari mana harus memulai. Anda akan melihat ide untuk mengubah alur kerja pengeditan menjadi alat MCP, meningkatkan halaman arahan, dan membandingkan dua pengiriman suara. Open source memberi Anda lebih banyak kendali, tetapi model, hosting, dan komputasi tetap membutuhkan biaya. Apa yang akan Anda pelajari: — Membangun ruang kerja rekan kerja AI dengan OpenDots: CopilotKit/OpenDots — Mengubah fungsi Python menjadi alat AI dengan FastMCP: PrefectHQ/fastmcp — Mengkritik dan menyempurnakan desain situs web dengan Impeccable: pbakaus/impeccable — Memeriksa perangkat lunak dan menciptakan kembali perilaku tertentu dengan REA: rea.tools — Membandingkan penyampaian suara dengan Chatterbox: resemble-ai/chatterbox Tonton selanjutnya: Codex Masterclass: youtu.be Claude Code Masterclass: youtu.be Membangun SaaS Pertama Anda: youtu.be Tautan Saya: Berkreasi dengan AI: skool.com Typeless, the voice Dikte yang saya gunakan setiap hari: typeless.com Skrip & kemas video Anda dengan AI: scripti.ai Berita teknologi mentah tanpa pelacakan: techl8.com Bab: 0:00 Lima alat AI sumber terbuka 0:29 OpenDots — rekan kerja AI 2:27 FastMCP — alat AI khusus 4:35 Arcade — sponsor 5:38 Impeccable — desain situs web yang lebih baik 7:23 REA — pahami cara kerja perangkat lunak 9:21 Chatterbox — ucapan sumber terbuka 9:58 Perbandingan suara Chatterbox 11:32 Rekap dan alat mana yang harus dicoba terlebih dahulu Video ini disponsori oleh Arcade. Kontak: luke@byrnemedia.group [it] 5 strumenti di intelligenza artificiale open-source gratuiti che vale davvero la pena provare Prova Arcade, il runtime per MCP: arcade.dev.plug.dev Alcuni degli strumenti di intelligenza artificiale più interessanti sono progetti che puoi scaricare, eseguire e modificare autonomamente. In questo video illustro cinque progetti open source: OpenDots per collaboratori IA, FastMCP per strumenti IA personalizzati, Impeccable per la progettazione di siti web, REA per comprendere il funzionamento del software e Chatterbox per il riconoscimento vocale. Spiego cosa fa ciascuno di essi, fornisco un esempio utile e indico da dove iniziare. Vedrai idee per trasformare un flusso di lavoro di editing in uno strumento MCP, migliorare una landing page e confrontare due interpretazioni vocali. L'open source offre maggiore controllo, ma i modelli, l'hosting e la potenza di calcolo possono comunque avere un costo. Cosa imparerai: — Creare uno spazio di lavoro con un collaboratore AI con OpenDots: CopilotKit/OpenDots — Trasformare funzioni Python in strumenti AI con FastMCP: PrefectHQ/fastmcp — Analizzare e perfezionare il design di siti web con Impeccable: pbakaus/impeccable — Ispezionare software e ricreare un comportamento specifico con REA: rea.tools — Confrontare la pronuncia con Chatterbox: resemble-ai/chatterbox Guarda anche: Masterclass Codex: youtu.be Masterclass Claude Code: youtu.be Crea il tuo primo SaaS: youtu.be I miei link: Crea con l'IA: Italiano: skool.com Typeless, il sistema di dettatura vocale che uso ogni giorno: typeless.com Crea script e confeziona i tuoi video con l'IA: scripti.ai Notizie tecnologiche pure e senza tracciamento: techl8.com Capitoli: 0:00 Cinque strumenti open source basati sull'IA 0:29 OpenDots: collaboratori IA 2:27 FastMCP: strumenti IA personalizzati 4:35 Arcade: sponsor 5:38 Impeccable: design del sito web migliore 7:23 REA: capire come funziona un software 9:21 Chatterbox: sintesi vocale open source 9:58 Confronto tra le voci di Chatterbox 11:32 Riepilogo e quale strumento provare per primo Questo video è sponsorizzato da Arcade. Contatti: luke@byrnemedia.group [iw] 5 כלי AI בקוד פתוח בחינם שבאמת שווה לנסות נסו את Arcade, זמן הריצה של MCP: arcade.dev.plug.dev חלק מכלי הבינה המלאכותית המעניינים ביותר הם פרויקטים שתוכלו להוריד, להפעיל ולשנות בעצמכם. בסרטון זה אני עובר על חמישה פרויקטים בקוד פתוח: OpenDots לעמיתים לעבודה בתחום הבינה המלאכותית, FastMCP לכלי בינה מלאכותית מותאמים אישית, Impeccable לעיצוב אתרים, REA להבנת אופן פעולת התוכנה ו-Chatterbox לדיבור. אני מכסה מה כל אחד מהם עושה, דוגמה שימושית, והיכן להתחיל. תראו רעיונות להפיכת זרימת עבודה של עריכה לכלי MCP, שיפור דף נחיתה והשוואה בין שני סוגי העברת קול. קוד פתוח נותן לכם יותר שליטה, אך מודלים, אירוח ומחשוב עדיין יכולים לעלות כסף. מה תלמדו: — בניית סביבת עבודה לעבודה מבוססת בינה מלאכותית עם OpenDots: CopilotKit/OpenDots — הפיכת פונקציות Python לכלי בינה מלאכותית עם FastMCP: PrefectHQ/fastmcp — ביקורת וליטוש של עיצובי אתרים עם Impeccable: pbakaus/impeccable — בדיקת תוכנה ושחזור התנהגות ספציפית עם REA: rea.tools — השווה דיבור עם Chatterbox: resemble-ai/chatterbox צפו בהמשך: כיתת אמן של Codex: youtu.be כיתת אמן של Claude Code: youtu.be בניית ה-SaaS הראשון שלך: youtu.be הקישורים שלי: יצירה עם בינה מלאכותית: skool.com Typeless, ההכתבה הקולית שאני משתמש בה בכל יום: typeless.com כתבו וארוזו את הסרטונים שלכם עם בינה מלאכותית: scripti.ai חדשות טכנולוגיה גולמיות ללא מעקב: techl8.com פרקים: 0:00 חמישה כלי בינה מלאכותית בקוד פתוח 0:29 OpenDots - עמיתים לעבודה בתחום הבינה המלאכותית 2:27 FastMCP - כלי בינה מלאכותית מותאמים אישית 4:35 Arcade - ספונסר 5:38 Impeccable - עיצוב אתר טוב יותר 7:23 REA - הבנת איך תוכנה עובדת 9:21 Chatterbox - דיבור בקוד פתוח 9:58 השוואת קול Chatterbox 11:32 סיכום ואיזה כלי לנסות קודם סרטון זה ממומן על ידי Arcade. איש קשר: luke@byrnemedia.group [ja] 試してみる価値のある無料のオープンソースAIツール5選 MCPのランタイムであるArcadeをお試しください：arcade.dev.plug.dev 最も興味深いAIツールのいくつかは、ダウンロードして実行し、自分で変更できるプロジェクトです。このビデオでは、5つのオープンソースプロジェクトを紹介します。AI同僚のためのOpenDots、カスタムAIツールのためのFastMCP、WebサイトデザインのためのImpeccable、ソフトウェアの仕組みを理解するためのREA、そして音声認識のためのChatterboxです。それぞれの機能、役立つ例、そして始め方について説明します。編集ワークフローをMCPツールに変換するアイデア、ランディングページを改善する方法、2つの音声配信を比較する方法なども紹介します。オープンソースはより多くの制御を可能にしますが、モデル、ホスティング、コンピューティングには費用がかかる場合があります。 学習内容： — OpenDotsでAIコワーカーワークスペースを構築する：CopilotKit/OpenDots — FastMCPでPython関数をAIツールに変換する：PrefectHQ/fastmcp — Impeccableでウェブサイトのデザインを批評し、磨き上げる：pbakaus/impeccable — REAでソフトウェアを検査し、特定の動作を再現する：rea.tools — Chatterboxで音声の発話を比較する：resemble-ai/chatterbox 次の動画： Codexマスタークラス：youtu.be Claude Codeマスタークラス：youtu.be 初めてのSaaSを構築する：youtu.be 私のリンク： AIで創造する：skool.com Typeless私が毎日使っている音声認識ツール：typeless.com AIで動画のスクリプト作成とパッケージング：scripti.ai トラッキングなしの生のテクノロジーニュース：techl8.com チャプター： 0:00 5つのオープンソースAIツール 0:29 OpenDots ― AIの同僚 2:27 FastMCP ― カスタムAIツール 4:35 Arcade ― スポンサー 5:38 Impeccable ― より良いウェブサイトデザイン 7:23 REA ― ソフトウェアの仕組みを理解する 9:21 Chatterbox ― オープンソース音声認識ツール 9:58 Chatterboxの音声比較 11:32 まとめと最初に試すべきツール この動画はArcadeの提供でお送りします。 お問い合わせ：luke@byrnemedia.group [ko] 꼭 써봐야 할 무료 오픈소스 AI 도구 5가지 MCP 런타임인 Arcade를 사용해 보세요: arcade.dev.plug.dev 가장 흥미로운 AI 도구 중 일부는 직접 다운로드하고 실행하고 수정할 수 있는 프로젝트입니다. 이 영상에서는 AI 협업 도구인 OpenDots, 맞춤형 AI 도구인 FastMCP, 웹사이트 디자인 도구인 Impeccable, 소프트웨어 작동 방식 이해를 위한 REA, 그리고 음성 인식 도구인 Chatterbox 등 다섯 가지 오픈 소스 프로젝트를 소개합니다. 각 프로젝트의 기능, 유용한 예시, 그리고 시작하는 방법을 설명합니다. 편집 워크플로를 MCP 도구로 활용하는 방법, 랜딩 페이지를 개선하는 방법, 두 가지 음성 녹음을 비교하는 방법 등에 대한 아이디어를 얻을 수 있습니다. 오픈 소스는 더 많은 제어권을 제공하지만, 모델, 호스팅 및 컴퓨팅 비용은 여전히 ​​발생할 수 있습니다. 이 강의에서 배우게 될 내용: — OpenDots로 AI 협업 공간 구축하기: CopilotKit/OpenDots — FastMCP로 Python 함수를 AI 도구로 변환하기: PrefectHQ/fastmcp — Impeccable로 웹사이트 디자인 검토 및 개선하기: pbakaus/impeccable — REA로 소프트웨어 검사 및 특정 동작 재현하기: rea.tools — Chatterbox로 음성 전달 비교하기: resemble-ai/chatterbox 다음 강의: Codex 마스터클래스: youtu.be Claude Code 마스터클래스: youtu.be 첫 번째 SaaS 구축하기: youtu.be 내 링크: AI로 만들기: skool.com 제가 매일 사용하는 음성 인식 도구 Typeless: typeless.com AI로 영상 스크립트 작성 및 패키징: scripti.ai 추적 없는 생생한 기술 뉴스: techl8.com 목차: 0:00 5가지 오픈 소스 AI 도구 0:29 OpenDots — AI 협업 도구 2:27 FastMCP — 맞춤형 AI 도구 4:35 Arcade — 스폰서 5:38 Impeccable — 더 나은 웹사이트 디자인 7:23 REA — 소프트웨어 작동 방식 이해 9:21 Chatterbox — 오픈 소스 음성 인식 9:58 Chatterbox 음성 비교 11:32 요약 및 먼저 사용해 볼 도구 이 영상은 Arcade의 후원을 받았습니다. 문의: luke@byrnemedia.group [ml] ശ്രമിക്കേണ്ട 5 സൗജന്യ ഓപ്പൺ സോഴ്‌സ് AI ഉപകരണങ്ങൾ MCP-യുടെ റൺടൈമായ ആർക്കേഡ് പരീക്ഷിച്ചുനോക്കൂ: arcade.dev.plug.dev ഏറ്റവും രസകരമായ ചില AI ടൂളുകൾ നിങ്ങൾക്ക് ഡൗൺലോഡ് ചെയ്യാനും പ്രവർത്തിപ്പിക്കാനും സ്വയം മാറ്റാനും കഴിയുന്ന പ്രോജക്റ്റുകളാണ്. ഈ വീഡിയോയിൽ ഞാൻ അഞ്ച് ഓപ്പൺ സോഴ്‌സ് പ്രോജക്റ്റുകളിലൂടെ കടന്നുപോകുന്നു: AI സഹപ്രവർത്തകർക്കുള്ള OpenDots, ഇഷ്ടാനുസൃത AI ടൂളുകൾക്കുള്ള FastMCP, വെബ്‌സൈറ്റ് ഡിസൈനിന് Impeccable, സോഫ്റ്റ്‌വെയർ എങ്ങനെ പ്രവർത്തിക്കുന്നുവെന്ന് മനസ്സിലാക്കാൻ REA, സംഭാഷണത്തിനുള്ള Chatterbox. ഓരോന്നും എന്താണ് ചെയ്യുന്നതെന്ന്, ഒരു ഉപയോഗപ്രദമായ ഉദാഹരണം, എവിടെ തുടങ്ങണം എന്നിവ ഞാൻ ഉൾക്കൊള്ളുന്നു. ഒരു എഡിറ്റിംഗ് വർക്ക്ഫ്ലോയെ ഒരു MCP ടൂളാക്കി മാറ്റുന്നതിനും, ഒരു ലാൻഡിംഗ് പേജ് മെച്ചപ്പെടുത്തുന്നതിനും, രണ്ട് വോയ്‌സ് ഡെലിവറികൾ താരതമ്യം ചെയ്യുന്നതിനുമുള്ള ആശയങ്ങൾ നിങ്ങൾ കാണും. ഓപ്പൺ സോഴ്‌സ് നിങ്ങൾക്ക് കൂടുതൽ നിയന്ത്രണം നൽകുന്നു, പക്ഷേ മോഡലുകൾ, ഹോസ്റ്റിംഗ്, കമ്പ്യൂട്ട് എന്നിവ ഇപ്പോഴും പണം ചിലവാകും. നിങ്ങൾക്ക് പഠിക്കാൻ കഴിയുന്നത്: — OpenDots ഉപയോഗിച്ച് ഒരു AI സഹപ്രവർത്തക വർക്ക്‌സ്‌പെയ്‌സ് നിർമ്മിക്കുക: CopilotKit/OpenDots — FastMCP ഉപയോഗിച്ച് പൈത്തൺ ഫംഗ്‌ഷനുകളെ AI ടൂളുകളാക്കി മാറ്റുക: PrefectHQ/fastmcp — Impeccable ഉപയോഗിച്ച് വെബ്‌സൈറ്റ് ഡിസൈനുകളെ വിമർശിക്കുകയും പോളിഷ് ചെയ്യുകയും ചെയ്യുക: pbakaus/impeccable — REA ഉപയോഗിച്ച് സോഫ്റ്റ്‌വെയർ പരിശോധിക്കുകയും ഒരു പ്രത്യേക പെരുമാറ്റം പുനഃസൃഷ്ടിക്കുകയും ചെയ്യുക: rea.tools — Chatterbox ഉപയോഗിച്ച് സംഭാഷണ ഡെലിവറി താരതമ്യം ചെയ്യുക: resemble-ai/chatterbox അടുത്തത് കാണുക: Codex Masterclass: youtu.be Claude Code Masterclass: youtu.be നിങ്ങളുടെ ആദ്യ SaaS നിർമ്മിക്കുക: youtu.be എന്റെ ലിങ്കുകൾ: AI ഉപയോഗിച്ച് സൃഷ്ടിക്കുക: skool.com ടൈപ്പ്‌ലെസ്, ശബ്ദം ഞാൻ എല്ലാ ദിവസവും ഉപയോഗിക്കുന്ന ഡിക്റ്റേഷൻ: typeless.com നിങ്ങളുടെ വീഡിയോകൾ AI ഉപയോഗിച്ച് സ്ക്രിപ്റ്റ് ചെയ്ത് പാക്കേജ് ചെയ്യുക: scripti.ai ട്രാക്കിംഗ് ഇല്ലാതെ റോ ടെക് വാർത്തകൾ: techl8.com അധ്യായങ്ങൾ: 0:00 അഞ്ച് ഓപ്പൺ സോഴ്‌സ് AI ഉപകരണങ്ങൾ 0:29 ഓപ്പൺഡോട്ടുകൾ — AI സഹപ്രവർത്തകർ 2:27 FastMCP — ഇഷ്ടാനുസൃത AI ഉപകരണങ്ങൾ 4:35 ആർക്കേഡ് — സ്പോൺസർ 5:38 Impecable — മികച്ച വെബ്‌സൈറ്റ് ഡിസൈൻ 7:23 REA — സോഫ്റ്റ്‌വെയർ എങ്ങനെ പ്രവർത്തിക്കുന്നുവെന്ന് മനസ്സിലാക്കുക 9:21 ചാറ്റർബോക്സ് — ഓപ്പൺ സോഴ്‌സ് പ്രസംഗം 9:58 ചാറ്റർബോക്സ് വോയ്‌സ് താരതമ്യം 11:32 റീക്യാപ്പും ആദ്യം ഏത് ടൂളും പരീക്ഷിക്കണം ഈ വീഡിയോ ആർക്കേഡ് സ്പോൺസർ ചെയ്‌തതാണ്. ബന്ധപ്പെടുക: luke@byrnemedia.group [nl-NL] 5 gratis open-source AI-tools die het echt waard zijn om te proberen Probeer Arcade, de runtime voor MCP: arcade.dev.plug.dev Sommige van de meest interessante AI-tools zijn projecten die je kunt downloaden, uitvoeren en zelf aanpassen. In deze video bespreek ik vijf open-sourceprojecten: OpenDots voor AI-collega's, FastMCP voor aangepaste AI-tools, Impeccable voor webdesign, REA om te begrijpen hoe software werkt en Chatterbox voor spraakherkenning. Ik leg uit wat elk project doet, geef een nuttig voorbeeld en vertel waar je kunt beginnen. Je ziet ideeën om een ​​bewerkingsworkflow om te zetten in een MCP-tool, een landingspagina te verbeteren en twee spraakopnames te vergelijken. Open source geeft je meer controle, maar modellen, hosting en rekenkracht kunnen nog steeds geld kosten. Wat je leert: — Bouw een AI-collega-werkruimte met OpenDots: CopilotKit/OpenDots — Zet Python-functies om in AI-tools met FastMCP: PrefectHQ/fastmcp — Beoordeel en perfectioneer websiteontwerpen met Impeccable: pbakaus/impeccable — Inspecteer software en reproduceer specifiek gedrag met REA: rea.tools — Vergelijk spraak met Chatterbox: resemble-ai/chatterbox Bekijk ook: Codex Masterclass: youtu.be Claude Code Masterclass: youtu.be Bouw je eerste SaaS: youtu.be Mijn links: Creëer met AI: skool.com Typeless, the Spraakdictatie die ik dagelijks gebruik: typeless.com Script en verpak je video's met AI: scripti.ai Onbewerkt technieuws zonder tracking: techl8.com Hoofdstukken: 0:00 Vijf open-source AI-tools 0:29 OpenDots — AI-collega's 2:27 FastMCP — aangepaste AI-tools 4:35 Arcade — sponsor 5:38 Impeccable — beter websiteontwerp 7:23 REA — begrijpen hoe software werkt 9:21 Chatterbox — open-source spraakherkenning 9:58 Chatterbox-stemvergelijking 11:32 Samenvatting en welke tool je als eerste moet proberen Deze video is gesponsord door Arcade. Contact: luke@byrnemedia.group [pa] 5 ਮੁਫ਼ਤ ਓਪਨ-ਸੋਰਸ AI ਟੂਲ ਜੋ ਅਸਲ ਵਿੱਚ ਵਰਤਣ ਯੋਗ ਹਨ MCP ਲਈ ਰਨਟਾਈਮ, ਆਰਕੇਡ ਅਜ਼ਮਾਓ: arcade.dev.plug.dev ਕੁਝ ਸਭ ਤੋਂ ਦਿਲਚਸਪ AI ਟੂਲ ਉਹ ਪ੍ਰੋਜੈਕਟ ਹਨ ਜਿਨ੍ਹਾਂ ਨੂੰ ਤੁਸੀਂ ਡਾਊਨਲੋਡ ਕਰ ਸਕਦੇ ਹੋ, ਚਲਾ ਸਕਦੇ ਹੋ ਅਤੇ ਆਪਣੇ ਆਪ ਨੂੰ ਬਦਲ ਸਕਦੇ ਹੋ। ਇਸ ਵੀਡੀਓ ਵਿੱਚ ਮੈਂ ਪੰਜ ਓਪਨ-ਸੋਰਸ ਪ੍ਰੋਜੈਕਟਾਂ ਵਿੱਚੋਂ ਲੰਘਦਾ ਹਾਂ: AI ਸਹਿਕਰਮੀਆਂ ਲਈ ਓਪਨਡੌਟਸ, ਕਸਟਮ AI ਟੂਲਸ ਲਈ ਫਾਸਟਐਮਸੀਪੀ, ਵੈੱਬਸਾਈਟ ਡਿਜ਼ਾਈਨ ਲਈ ਇਮਪੀਕੇਬਲ, ਸੌਫਟਵੇਅਰ ਕਿਵੇਂ ਕੰਮ ਕਰਦਾ ਹੈ ਇਹ ਸਮਝਣ ਲਈ REA, ਅਤੇ ਭਾਸ਼ਣ ਲਈ ਚੈਟਰਬਾਕਸ। ਮੈਂ ਇਹ ਕਵਰ ਕਰਦਾ ਹਾਂ ਕਿ ਹਰ ਇੱਕ ਕੀ ਕਰਦਾ ਹੈ, ਇੱਕ ਉਪਯੋਗੀ ਉਦਾਹਰਣ, ਅਤੇ ਕਿੱਥੋਂ ਸ਼ੁਰੂ ਕਰਨਾ ਹੈ। ਤੁਸੀਂ ਇੱਕ ਸੰਪਾਦਨ ਵਰਕਫਲੋ ਨੂੰ ਇੱਕ MCP ਟੂਲ ਵਿੱਚ ਬਦਲਣ, ਇੱਕ ਲੈਂਡਿੰਗ ਪੇਜ ਨੂੰ ਬਿਹਤਰ ਬਣਾਉਣ ਅਤੇ ਦੋ ਵੌਇਸ ਡਿਲੀਵਰੀਆਂ ਦੀ ਤੁਲਨਾ ਕਰਨ ਲਈ ਵਿਚਾਰ ਵੇਖੋਗੇ। ਓਪਨ ਸੋਰਸ ਤੁਹਾਨੂੰ ਵਧੇਰੇ ਨਿਯੰਤਰਣ ਦਿੰਦਾ ਹੈ, ਪਰ ਮਾਡਲ, ਹੋਸਟਿੰਗ ਅਤੇ ਕੰਪਿਊਟ ਅਜੇ ਵੀ ਪੈਸੇ ਖਰਚ ਕਰ ਸਕਦੇ ਹਨ। ਤੁਸੀਂ ਕੀ ਸਿੱਖੋਗੇ: — OpenDots ਨਾਲ ਇੱਕ AI ਸਹਿ-ਕਰਮਚਾਰੀ ਵਰਕਸਪੇਸ ਬਣਾਓ: CopilotKit/OpenDots — FastMCP ਨਾਲ Python ਫੰਕਸ਼ਨਾਂ ਨੂੰ AI ਟੂਲਸ ਵਿੱਚ ਬਦਲੋ: PrefectHQ/fastmcp — Impeccable ਨਾਲ ਵੈੱਬਸਾਈਟ ਡਿਜ਼ਾਈਨ ਦੀ ਆਲੋਚਨਾ ਅਤੇ ਪਾਲਿਸ਼ ਕਰੋ: pbakaus/impeccable — ਸਾਫਟਵੇਅਰ ਦਾ ਨਿਰੀਖਣ ਕਰੋ ਅਤੇ REA ਨਾਲ ਇੱਕ ਖਾਸ ਵਿਵਹਾਰ ਨੂੰ ਮੁੜ ਬਣਾਓ: rea.tools — ਚੈਟਰਬਾਕਸ ਨਾਲ ਸਪੀਚ ਡਿਲੀਵਰੀ ਦੀ ਤੁਲਨਾ ਕਰੋ: resemble-ai/chatterbox ਅੱਗੇ ਦੇਖੋ: ਕੋਡੈਕਸ ਮਾਸਟਰਕਲਾਸ: youtu.be ਕਲਾਉਡ ਕੋਡ ਮਾਸਟਰਕਲਾਸ: youtu.be ਆਪਣਾ ਪਹਿਲਾ SaaS ਬਣਾਓ: youtu.be ਮੇਰੇ ਲਿੰਕ: AI ਨਾਲ ਬਣਾਓ: skool.com ਟਾਈਪਲੈੱਸ, ਆਵਾਜ਼ ਡਿਕਟੇਸ਼ਨ ਜੋ ਮੈਂ ਹਰ ਰੋਜ਼ ਵਰਤਦਾ ਹਾਂ: typeless.com AI ਨਾਲ ਆਪਣੇ ਵੀਡੀਓਜ਼ ਨੂੰ ਸਕ੍ਰਿਪਟ ਅਤੇ ਪੈਕੇਜ ਕਰੋ: scripti.ai ਕੋਈ ਟਰੈਕਿੰਗ ਤੋਂ ਬਿਨਾਂ ਕੱਚੀ ਤਕਨੀਕੀ ਖ਼ਬਰਾਂ: techl8.com ਅਧਿਆਇ: 0:00 ਪੰਜ ਓਪਨ-ਸੋਰਸ AI ਟੂਲ 0:29 ਓਪਨਡੌਟਸ — AI ਸਹਿਕਰਮੀ 2:27 ਫਾਸਟਐਮਸੀਪੀ — ਕਸਟਮ AI ਟੂਲ 4:35 ਆਰਕੇਡ — ਸਪਾਂਸਰ 5:38 ਨਿਰਦੋਸ਼ — ਬਿਹਤਰ ਵੈੱਬਸਾਈਟ ਡਿਜ਼ਾਈਨ 7:23 REA — ਸਮਝੋ ਕਿ ਸਾਫਟਵੇਅਰ ਕਿਵੇਂ ਕੰਮ ਕਰਦਾ ਹੈ 9:21 ਚੈਟਰਬਾਕਸ — ਓਪਨ-ਸੋਰਸ ਭਾਸ਼ਣ 9:58 ਚੈਟਰਬਾਕਸ ਵੌਇਸ ਤੁਲਨਾ 11:32 ਰੀਕੈਪ ਅਤੇ ਪਹਿਲਾਂ ਕਿਹੜਾ ਟੂਲ ਅਜ਼ਮਾਉਣਾ ਹੈ ਇਹ ਵੀਡੀਓ ਆਰਕੇਡ ਦੁਆਰਾ ਸਪਾਂਸਰ ਕੀਤਾ ਗਿਆ ਹੈ। ਸੰਪਰਕ: luke@byrnemedia.group [pl] 5 darmowych narzędzi AI typu open-source, które naprawdę warto wypróbować Wypróbuj Arcade, środowisko uruchomieniowe dla MCP: arcade.dev.plug.dev Niektóre z najciekawszych narzędzi AI to projekty, które można pobrać, uruchomić i samodzielnie modyfikować. W tym filmie omawiam pięć projektów open source: OpenDots dla współpracowników zajmujących się AI, FastMCP dla niestandardowych narzędzi AI, Impeccable do projektowania stron internetowych, REA do zrozumienia działania oprogramowania oraz Chatterbox do obsługi mowy. Omawiam, do czego służy każdy z nich, podaję przydatny przykład i od czego zacząć. Zobaczysz pomysły na przekształcenie procesu edycji w narzędzie MCP, ulepszenie strony docelowej i porównanie dwóch wersji głosowych. Oprogramowanie open source daje większą kontrolę, ale modele, hosting i obliczenia nadal mogą kosztować. Czego się nauczysz: — Zbuduj przestrzeń roboczą AI dla współpracowników za pomocą OpenDots: CopilotKit/OpenDots — Przekształć funkcje Pythona w narzędzia AI za pomocą FastMCP: PrefectHQ/fastmcp — Krytykuj i udoskonalaj projekty stron internetowych za pomocą Impeccable: pbakaus/impeccable — Inspekcja oprogramowania i odtworzenie określonego zachowania za pomocą REA: rea.tools — Porównaj sposób przekazu mowy z Chatterbox: resemble-ai/chatterbox Obejrzyj dalej: Kurs mistrzowski Codex: youtu.be Kurs mistrzowski Claude Code: youtu.be Zbuduj swój pierwszy SaaS: youtu.be Mój Linki: Twórz z AI: skool.com Typeless, dyktafon, którego używam na co dzień: typeless.com Twórz scenariusze i pakuj filmy z AI: scripti.ai Surowe wiadomości technologiczne bez śledzenia: techl8.com Rozdziały: 0:00 Pięć narzędzi AI typu open source 0:29 OpenDots — współpracownicy AI 2:27 FastMCP — niestandardowe narzędzia AI 4:35 Arcade — sponsor 5:38 Impeccable — lepszy projekt strony internetowej 7:23 REA — zrozum, jak działa oprogramowanie 9:21 Chatterbox — mowa typu open source 9:58 Porównanie głosu Chatterbox 11:32 Podsumowanie i które narzędzie wypróbować najpierw Ten film jest sponsorowany przez Arcade. Kontakt: luke@byrnemedia.group [pt-BR] 5 ferramentas de IA open-source gratuitas que realmente valem a pena testar Experimente o Arcade, o ambiente de execução para MCP: arcade.dev.plug.dev Algumas das ferramentas de IA mais interessantes são projetos que você pode baixar, executar e modificar. Neste vídeo, apresento cinco projetos de código aberto: OpenDots para IA colaborativa, FastMCP para ferramentas de IA personalizadas, Impeccable para design de sites, REA para entender como o software funciona e Chatterbox para reconhecimento de voz. Explico o que cada um faz, apresento um exemplo útil e explico por onde começar. Você verá ideias para transformar um fluxo de trabalho de edição em uma ferramenta MCP, aprimorar uma landing page e comparar duas versões de voz. O código aberto oferece mais controle, mas modelos, hospedagem e computação ainda podem ter custos. O que você aprenderá: — Criar um espaço de trabalho colaborativo com IA usando o OpenDots: CopilotKit/OpenDots — Transformar funções Python em ferramentas de IA com o FastMCP: PrefectHQ/fastmcp — Analisar e aprimorar designs de sites com o Impeccable: pbakaus/impeccable — Inspecionar softwares e recriar comportamentos específicos com o REA: rea.tools — Comparar a entrega de fala com o Chatterbox: resemble-ai/chatterbox Assista a seguir: Masterclass Codex: youtu.be Masterclass Claude Code: youtu.be Crie seu primeiro SaaS: youtu.be Meus links: Crie com IA: skool.com Typeless, a voz Ditado que uso diariamente: typeless.com Roteirize e empacote seus vídeos com IA: scripti.ai Notícias de tecnologia sem rastreamento: techl8.com Capítulos: 0:00 Cinco ferramentas de IA de código aberto 0:29 OpenDots — colegas de trabalho de IA 2:27 FastMCP — ferramentas de IA personalizadas 4:35 Arcade — patrocinador 5:38 Impeccable — design de sites aprimorado 7:23 REA — entenda como o software funciona 9:21 Chatterbox — reconhecimento de voz de código aberto 9:58 Comparação de vozes do Chatterbox 11:32 Resumo e qual ferramenta experimentar primeiro Este vídeo é patrocinado pela Arcade. Contato: luke@byrnemedia.group [ru] 5 бесплатных ИИ-инструментов с открытым кодом, которые действительно стоит попробовать Попробуйте Arcade, среду выполнения для MCP: arcade.dev.plug.dev Некоторые из самых интересных инструментов ИИ — это проекты, которые вы можете скачать, запустить и изменить самостоятельно. В этом видео я расскажу о пяти проектах с открытым исходным кодом: OpenDots для коллег-ИИ, FastMCP для создания собственных инструментов ИИ, Impeccable для веб-дизайна, REA для понимания принципов работы программного обеспечения и Chatterbox для распознавания речи. Я расскажу о том, что делает каждый из них, приведу полезный пример и подскажу, с чего начать. Вы увидите идеи по превращению рабочего процесса редактирования в инструмент MCP, улучшению целевой страницы и сравнению двух вариантов озвучивания. Открытый исходный код дает вам больше контроля, но модели, хостинг и вычислительные ресурсы все еще могут стоить денег. Что вы узнаете: — Создайте рабочее пространство для коллег с ИИ с помощью OpenDots: CopilotKit/OpenDots — Превратите функции Python в инструменты ИИ с помощью FastMCP: PrefectHQ/fastmcp — Критикуйте и улучшайте дизайн веб-сайтов с помощью Impeccable: pbakaus/impeccable — Проверяйте программное обеспечение и воссоздавайте определенное поведение с помощью REA: rea.tools — Сравнивайте качество речи с Chatterbox: resemble-ai/chatterbox Смотрите дальше: Мастер-класс Codex: youtu.be Мастер-класс Claude Code: youtu.be Создайте свой первый SaaS: youtu.be Мои ссылки: Создавайте с помощью ИИ: skool.com Typeless, голосовой диктофон, который я использую каждый день: typeless.com Создавайте сценарии и упаковывайте свои видео с помощью ИИ: scripti.ai Необработанные новости технологий без отслеживания: techl8.com Главы: 0:00 Пять инструментов ИИ с открытым исходным кодом 0:29 OpenDots — коллеги с ИИ 2:27 FastMCP — пользовательские инструменты ИИ 4:35 Arcade — спонсор 5:38 Impeccable — улучшенный дизайн веб-сайтов 7:23 REA — понимание принципов работы программного обеспечения 9:21 Chatterbox — голосовой диктофон с открытым исходным кодом 9:58 Сравнение голосов Chatterbox 11:32 Итоги и какой инструмент попробовать первым Это видео спонсируется компанией Arcade. Контакты: luke@byrnemedia.group [ta] நிச்சயமாக முயற்சி செய்ய வேண்டிய 5 இலவச ஓப்பன்-சோர்ஸ் AI கருவிகள் MCP-க்கான ரன்டைமான ஆர்கேடை (Arcade) முயற்சித்துப் பாருங்கள்: arcade.dev.plug.dev மிகவும் சுவாரஸ்யமான சில AI கருவிகள், நீங்களே பதிவிறக்கம் செய்து, இயக்கி, மாற்றியமைக்கக்கூடிய திட்டங்களாகும். இந்த வீடியோவில், நான் ஐந்து ஓப்பன்-சோர்ஸ் திட்டங்களைப் பற்றி விளக்குகிறேன்: AI உடன் பணிபுரிபவர்களுக்கான ஓப்பன்டாட்ஸ் (OpenDots), தனிப்பயன் AI கருவிகளுக்கான ஃபாஸ்ட்எம்சிபி (FastMCP), இணையதள வடிவமைப்பிற்கான இம்பெக்கபிள் (Impeccable), மென்பொருள் எவ்வாறு செயல்படுகிறது என்பதைப் புரிந்துகொள்வதற்கான ஆர்இஏ (REA), மற்றும் பேச்சுக்கான சாட்டர்பாக்ஸ் (Chatterbox). ஒவ்வொன்றும் என்ன செய்கிறது, ஒரு பயனுள்ள உதாரணம், மற்றும் எங்கிருந்து தொடங்குவது என்பதை நான் விளக்குகிறேன். ஒரு எடிட்டிங் பணிப்பாய்வை (workflow) MCP கருவியாக மாற்றுவது, ஒரு லேண்டிங் பக்கத்தை மேம்படுத்துவது, மற்றும் இரண்டு குரல் வழங்கல்களை ஒப்பிடுவது போன்ற யோசனைகளை நீங்கள் காண்பீர்கள். ஓப்பன் சோர்ஸ் உங்களுக்கு அதிகக் கட்டுப்பாட்டை வழங்குகிறது, ஆனால் மாடல்கள், ஹோஸ்டிங் மற்றும் கம்ப்யூட் ஆகியவற்றுக்கு இன்னும் பணம் செலவாகலாம். நீங்கள் கற்றுக்கொள்வது: — OpenDots-ஐப் பயன்படுத்தி ஒரு AI சக பணியாளர் பணிச்சூழலை உருவாக்குங்கள்: CopilotKit/OpenDots — FastMCP-ஐப் பயன்படுத்தி பைதான் செயல்பாடுகளை AI கருவிகளாக மாற்றுங்கள்: PrefectHQ/fastmcp — Impeccable-ஐப் பயன்படுத்தி இணையதள வடிவமைப்புகளை விமர்சித்து மெருகூட்டுங்கள்: pbakaus/impeccable — REA-ஐப் பயன்படுத்தி மென்பொருளை ஆய்வு செய்து ஒரு குறிப்பிட்ட நடத்தையை மீண்டும் உருவாக்குங்கள்: rea.tools — Chatterbox-ஐப் பயன்படுத்தி பேச்சு வழங்கலை ஒப்பிடுங்கள்: resemble-ai/chatterbox அடுத்துப் பாருங்கள்: Codex Masterclass: youtu.be Claude Code Masterclass: youtu.be உங்கள் முதல் SaaS-ஐ உருவாக்குங்கள்: youtu.be எனது இணைப்புகள்: AI உடன் உருவாக்குங்கள்: skool.com Typeless, நான் தினமும் பயன்படுத்தும் குரல்வழிச் சொல்லாக்கம் நாள்: typeless.com AI உடன் உங்கள் வீடியோக்களை ஸ்கிரிப்ட் செய்து தொகுக்கவும்: scripti.ai கண்காணிப்பு இல்லாத உண்மையான தொழில்நுட்பச் செய்திகள்: techl8.com அத்தியாயங்கள்: 0:00 ஐந்து திறந்த மூல AI கருவிகள் 0:29 OpenDots — AI உடன் பணிபுரிபவர்கள் 2:27 FastMCP — தனிப்பயன் AI கருவிகள் 4:35 Arcade — ஆதரவாளர் 5:38 Impeccable — சிறந்த இணையதள வடிவமைப்பு 7:23 REA — மென்பொருள் எவ்வாறு செயல்படுகிறது என்பதைப் புரிந்துகொள்ளுங்கள் 9:21 Chatterbox — திறந்த மூலப் பேச்சு 9:58 Chatterbox குரல் ஒப்பீடு 11:32 மீள்பார்வை மற்றும் முதலில் எந்தக் கருவியை முயற்சிப்பது இந்த வீடியோ Arcade நிறுவனத்தால் ஆதரிக்கப்படுகிறது. தொடர்புக்கு: luke@byrnemedia.group [te] ఖచ్చితంగా ప్రయత్నించదగ్గ 5 ఉచిత ఓపెన్-సోర్స్ AI టూల్స్ MCP కోసం రన్‌టైమ్ అయిన ఆర్కేడ్‌ను ప్రయత్నించండి: arcade.dev.plug.dev అత్యంత ఆసక్తికరమైన AI సాధనాలలో కొన్ని, మీరు స్వయంగా డౌన్‌లోడ్ చేసుకొని, రన్ చేసి, మార్పులు చేయగల ప్రాజెక్టులే. ఈ వీడియోలో నేను ఐదు ఓపెన్-సోర్స్ ప్రాజెక్టుల గురించి వివరిస్తాను: AI సహోద్యోగుల కోసం ఓపెన్‌డాట్స్, కస్టమ్ AI సాధనాల కోసం ఫాస్ట్‌ఎమ్‌సిపి, వెబ్‌సైట్ డిజైన్ కోసం ఇంపెక్కబుల్, సాఫ్ట్‌వేర్ ఎలా పనిచేస్తుందో అర్థం చేసుకోవడానికి ఆర్‌ఈఏ, మరియు స్పీచ్ కోసం చాటర్‌బాక్స్. వీటిలో ప్రతి ఒక్కటీ ఏమి చేస్తుందో, ఒక ఉపయోగకరమైన ఉదాహరణను, మరియు ఎక్కడ మొదలుపెట్టాలో నేను వివరిస్తాను. ఒక ఎడిటింగ్ వర్క్‌ఫ్లోను MCP సాధనంగా మార్చడం, ఒక ల్యాండింగ్ పేజీని మెరుగుపరచడం, మరియు రెండు వాయిస్ డెలివరీలను పోల్చడం వంటి ఆలోచనలను మీరు చూస్తారు. ఓపెన్ సోర్స్ మీకు మరింత నియంత్రణను ఇస్తుంది, కానీ మోడల్స్, హోస్టింగ్ మరియు కంప్యూట్‌కు ఇప్పటికీ డబ్బు ఖర్చు కావచ్చు. మీరు ఏమి నేర్చుకుంటారు: — OpenDots తో ఒక AI సహోద్యోగి వర్క్‌స్పేస్‌ను నిర్మించండి: CopilotKit/OpenDots — FastMCP తో పైథాన్ ఫంక్షన్‌లను AI సాధనాలుగా మార్చండి: PrefectHQ/fastmcp Impecable తో వెబ్‌సైట్ డిజైన్‌లను విమర్శించండి మరియు మెరుగుపరచండి: pbakaus/impeccable — REA తో సాఫ్ట్‌వేర్‌ను తనిఖీ చేయండి మరియు ఒక నిర్దిష్ట ప్రవర్తనను పునఃసృష్టించండి: rea.tools Chatterbox తో ప్రసంగ పంపిణీని పోల్చండి: resemble-ai/chatterbox తరువాత చూడండి: Codex మాస్టర్‌క్లాస్: youtu.be Claude కోడ్ మాస్టర్‌క్లాస్: youtu.be మీ మొదటి SaaS ను నిర్మించండి: youtu.be నా లింకులు: AI తో సృష్టించండి: skool.com Typeless, నేను ఉపయోగించే వాయిస్ డిక్టేషన్ ప్రతి రోజు: typeless.com AIతో మీ వీడియోలను స్క్రిప్ట్ చేయండి & ప్యాకేజ్ చేయండి: scripti.ai ట్రాకింగ్ లేని అసలైన టెక్ వార్తలు: techl8.com అధ్యాయాలు: 0:00 ఐదు ఓపెన్-సోర్స్ AI టూల్స్ 0:29 ఓపెన్‌డాట్స్ — AI సహోద్యోగులు 2:27 ఫాస్ట్‌ఎమ్‌సిపి — కస్టమ్ AI టూల్స్ 4:35 ఆర్కేడ్ — స్పాన్సర్ 5:38 ఇంపెక్కబుల్ — మెరుగైన వెబ్‌సైట్ డిజైన్ 7:23 ఆర్‌ఈఏ — సాఫ్ట్‌వేర్ ఎలా పనిచేస్తుందో అర్థం చేసుకోండి 9:21 చాటర్‌బాక్స్ — ఓపెన్-సోర్స్ స్పీచ్ 9:58 చాటర్‌బాక్స్ వాయిస్ పోలిక 11:32 రీక్యాప్ మరియు మొదట ఏ టూల్‌ను ప్రయత్నించాలి ఈ వీడియోను ఆర్కేడ్ స్పాన్సర్ చేసింది. సంప్రదించండి: luke@byrnemedia.group [uk] 5 безкоштовних інструментів штучного інтелекту з відкритим кодом, які дійсно варто спробувати Спробуйте Arcade, середовище виконання для MCP: arcade.dev.plug.dev Деякі з найцікавіших інструментів штучного інтелекту – це проекти, які ви можете завантажити, запустити та змінити самостійно. У цьому відео я розгляну п'ять проектів з відкритим кодом: OpenDots для колег зі штучного інтелекту, FastMCP для користувацьких інструментів штучного інтелекту, Impeccable для веб-дизайну, REA для розуміння того, як працює програмне забезпечення, та Chatterbox для мовлення. Я розгляну, що робить кожен з них, наведу корисний приклад і з чого почати. ​​Ви побачите ідеї для перетворення робочого процесу редагування на інструмент MCP, покращення цільової сторінки та порівняння двох голосових поставок. Відкритий код дає вам більше контролю, але моделі, хостинг та обчислення все ще можуть коштувати грошей. Чого ви навчитеся: — Створіть робочий простір для колег зі штучним інтелектом за допомогою OpenDots: CopilotKit/OpenDots — Перетворіть функції Python на інструменти штучного інтелекту за допомогою FastMCP: PrefectHQ/fastmcp — Критикуйте та вдосконалюйте дизайн веб-сайтів за допомогою Impeccable: pbakaus/impeccable — Перевіряйте програмне забезпечення та відтворюйте певну поведінку за допомогою REA: rea.tools — Порівнюйте передачу мовлення за допомогою Chatterbox: resemble-ai/chatterbox Дивіться далі: Майстер-клас Codex: youtu.be Майстер-клас Claude Code: youtu.be Створіть свій перший SaaS: youtu.be Мої посилання: Створюйте за допомогою штучного інтелекту: skool.com Типовий, голосовий диктант, який я використовую щодня день: typeless.com Створюйте сценарії та пакетуйте свої відео за допомогою ШІ: scripti.ai Сирі технологічні новини без відстеження: techl8.com Розділи: 0:00 П'ять інструментів ШІ з відкритим кодом 0:29 OpenDots — колеги зі ШІ 2:27 FastMCP — користувацькі інструменти ШІ 4:35 Arcade — спонсор 5:38 Impeccable — кращий дизайн веб-сайту 7:23 REA — розуміння того, як працює програмне забезпечення 9:21 Chatterbox — промова з відкритим кодом 9:58 Порівняння голосу Chatterbox 11:32 Підсумок та який інструмент спробувати першим Це відео спонсоровано Arcade. Контакт: luke@byrnemedia.group

A failure that looked like a capability gap between coding agents was caused by how the API key was passed Summary An AI agent compared three coding agents in a research repository that is scheduled to be made public. The research rebuilds an integrated agent development environment with OSS. The model and the tasks were the same for all three, and there were eight tasks. The source material consists of the records in this repository. In this comparison, a parent agent calls a parallel-execution MCP server. The parallel-execution MCP server is a program built as part of the research, and when it is called, it launches child agents in parallel. Only in the three runs with OpenHands CLI as the parent did none of the 32 child agents launched in total complete. Both the status field in the run record files and the result field of the measurement stayed normal, and for about a day this outcome was treated as a result of the comparison. The cause was not a difference in capability between the coding agents but the way the measurement script was built. The measurement script left the delivery of the API key that the child agents use to two things: inheritance of the parent process's environment variables and loading of a .env file. Only when the parent was OpenHands CLI and the run took place in an isolated working directory did the API key fail to arrive by either route. The AI agent fixed the measurement script so that it writes the API key in the env field of the MCP server configuration and passes it that way. After the fix, the AI agent measured the condition with OpenHands CLI as the parent again, once each in two cases: with the set of instruction files in place and without it. The set of instruction files is a group of files, such as AGENTS.md, placed in the working directory. Under both conditions, the API key no longer caused every child agent to fail. With the set of instruction files in place, all eight child agents completed. Without it, 15 of the 16 child agents launched did not complete, for reasons other than the API key. The first section of the body gives the overall picture of the cause and the fix. The sections after it are arranged in the following order. - Three sections explain how the comparison results were skewed, in this order: the measurement conditions and results, why the failure did not show up in the results, and how the cause was isolated - Whether the parent process's environment variables reach the MCP server is checked for each implementation in the public source code - Three sections explain the fix, in this order: the fix to how the API key is passed and the re-measurement, the mechanism that does not hide failures and the unit tests, and the protection of the API key - Before the section on the materials referred to, two rules for designing comparison measurements are presented The body names the products of the three coding agents compared. The content does not represent the views of any of their providers. What you can take away If you launch other agents from a coding agent through an MCP server, you will be able to do the following three things. - Check in the source code which of the parent process's environment variables your coding agent passes to a stdio MCP server. The range it passes differs by implementation - Fix your setup so that the API key the child agents use is written in the env field of the MCP server configuration and passed that way - Add to your own measurements a step that does not count failures caused by defects on the measurement side as results of a coding agent's capability This article is a discussion based on the run records in the research repository scheduled to be made public and on the public source code checked at the time of writing. --- Two implicit routes that the API key delivery depended on This section first explains why every child agent failed in the comparison measurements and what the setup looks like after the fix. The comparison measurements are a series of runs that measured three coding agents in turn, with the same model and the same tasks, while changing the conditions. A coding agent is a CLI program that is started from a terminal and that edits files and runs commands while calling a model. In the comparison measurements, the parent agent calls the parallel-execution MCP server. The parent agent is the coding agent process that receives the tasks and is started first. The parallel-execution MCP server is an MCP server built as part of the research, and when it is called, it launches child agents in parallel. An MCP server is a program that provides tools for a coding agent to call through a common protocol called MCP. The MCP servers in this article are of the stdio type: the agent starts them as child processes and communicates with them over standard input and output. A child agent is a coding agent process launched by the parallel-execution MCP server. The measurement script is a Python script that passes the tasks to the parent agent and starts it, then judges the results from the actual files and records them. Before the fix, the child agents' API key arrived by one of the following two implicit routes. The API key is the credential for calling the inference service. The inference service is a service that makes models available to call through an API. - Environment variable inheritance: the environment variables held by a parent process are passed as they are to the child processes it starts - Loading the .env file: the route by which the parallel-execution MCP server reads a .env file. A .env file holds configuration values such as API keys as lines of the form "name=value", and it is kept out of git Only when the parent was OpenHands CLI and the run also took place in an isolated working directory did the API key fail to reach the child agents by either of the two routes. An isolated working directory is a separate working directory of the repository, created with git worktree. This summary combines three conditions found in the research records, and the research records contain no statement in this form. The three conditions are explained in the section "Isolating the cause by rerunning with the run records kept". After the fix, the measurement script writes the API key in the env field of the MCP server configuration and passes it explicitly. The env field of the MCP server configuration is a field in the coding agent's configuration file. This field lists, for each MCP server, the environment variables to pass to the process that is started. The field is named environment in OpenCode and env in the other two. When the API key is needed but missing, the parallel-execution MCP server stops before launching child agents and marks the run record file as failed. A run record file is a JSON file that the parallel-execution MCP server writes for each call. It has fields for the number of child agents launched, the number of child agents completed, and the status. The following diagram shows the routes before and after the fix. Before the fix The shell that started the measurement script (with the API key loaded) └ Parent agent └ Parallel-execution MCP server │ Route 1: environment variable inheritance │ Did not arrive when the parent was OpenHands CLI │ Route 2: read the .env file at the top level of the repository │ The isolated working directory has no .env file └ Child agent (the API key value is an empty string) After the fix Measurement script └ Writes the API key in the env field of the MCP server configuration └ Parallel-execution MCP server ├ API key present: launches child agents └ API key needed but missing: stops before launching and marks the run record file as failed Conditions of the comparison measurements and the number of completed child agents per parent agent This section shows the conditions of the comparison measurements and the results for each parent agent in a table. The conditions of the comparison measurements were as follows. - Versions: OpenHands CLI version 1.16.0, Qwen Code version 0.21.13, and OpenCode version 1.18.18 were used - Model: one model was fixed for all runs and called through the OpenAI-compatible endpoint of the inference service - Tasks: eight independent programming tasks. Each task has a task statement and a test script - Instructions to the parent agent: two, namely to finish all eight without leaving any out and not to rewrite the test files. The instructions did not mention the name of the tool - How the parallel-execution MCP server was provided: it was registered in each coding agent's configuration file. Where it was registered differs among the three. For Qwen Code, it was registered in a configuration file placed in the working directory. For OpenCode, it was also registered in a configuration file placed in the working directory. For OpenHands CLI, it was registered in mcp.json placed in a dedicated storage directory - Child agents: for every parent agent, they were started in Qwen Code's non-interactive mode The following table summarizes, for each parent agent, the runs in which the parallel-execution MCP server was called. | Parent agent | Runs in which a call occurred | Run record files | Child agents launched | Child agents completed | |---|---|---|---|---| | Qwen Code as parent | 3 runs | 1, 1, and 2 per run | 8 per run | 8, 8, and 7 per run | | OpenCode as parent | 1 run | 3 | 16 in total | 8 in total | | OpenHands CLI as parent | 3 runs | 4 in total | 32 in total | 0 in total | In the three runs with OpenHands CLI as the parent, 0 of the 32 child agents launched completed. In the run with OpenCode as the parent, the parent agent fixed an error in the orchestration script by itself and ran it again. The orchestration script is a script, run by the parallel-execution MCP server, that describes the steps for launching child agents. Even in the three runs with OpenHands CLI as the parent, all eight tasks passed, and there were 0 rewrites of the test files. The records can be read as showing that the parent agent solved by itself the work that the child agents did not complete. However, the research records contain no breakdown of the parent agent's work. A judgment design in which the failure of every child agent did not show up in the results This section explains how a result in which every child agent failed continued to be treated as a result of the comparison. The measurement script judged whether the parallel-execution MCP server had been called by whether a run record file existed. Only the parallel-execution MCP server process writes run record files, so none is created if there is no call. The measurement script recorded the numbers of child agents launched and completed by summing the fields of the run record files. For pass or fail on the tasks, it restored the test files, ran the tests for each task, and recorded the number of tasks that passed. Even though every child agent failed, the failure did not show up in the results, for the following three reasons. - If a run record file existed, the measurement script treated the run as having a call and set the result field to success. Even if not a single child agent completed, the result field became success - The run record files written by the parallel-execution MCP server also all had ok, which indicates normal, in the status field - All eight tasks had passed, so the failure was not visible from the task results either The result in which OpenHands CLI was the parent and 0 child agents completed was treated as a result of the comparison measurements for about a day. Isolating the cause by rerunning with the run records kept This section explains how the cause was isolated, in the order of the work. According to the research records, the AI agent did the implementation, ran the measurements, and handled the PRs. This AI agent works separately from the reviewing AI agent that examines the diffs of PRs. The AI agent reran the condition with OpenHands CLI as the parent just once. This time, it kept the directory that holds the run record files instead of deleting it after the run. In the records it retrieved, all eight child agents had exited about 1.2 seconds after starting. The reason for exiting was an error saying that no authentication method had been selected. The text of this error is in the file in Qwen Code version 0.21.13 that checks authentication in non-interactive mode. When Qwen Code starts in non-interactive mode, it determines the authentication method from the command arguments, the environment variables, and the configuration files. If none of them determines it, Qwen Code stops with this error. The cause was the combination of the following three conditions. 1. When the parent was OpenHands CLI, the parent process's environment variables did not reach the parallel-execution MCP server process. When the parent was Qwen Code or OpenCode, they did reach it, so the API key loaded into the shell made it all the way to the child agents 2. When the API key was not in the environment variables, the parallel-execution MCP server was built to read the .env file in the top-level directory of its own repository. The comparison measurements ran in an isolated working directory, and that directory had no .env file 3. The measurement script did not write the API key in the env field of the MCP server configuration When launching child agents, the parallel-execution MCP server put an empty string into the environment variable if it could not get the API key value. Before that, there had also been a separate measurement with OpenHands CLI as the parent. In that measurement, the parent agent was explicitly made to call the parallel-execution MCP server, and one child agent had completed. The research records say that the explanation that fits best is that this run took place in the main working directory and the .env file was read. There is no record of which working directory it ran in, so this explanation has not been confirmed. Differences between implementations in whether the parent process's environment variables reach the MCP server This section shows the results of checking, for each implementation in the public source code, whether the parent process's environment variables reach a stdio MCP server. The checks were made on 2026-09-29. The following table summarizes the environment variables that each implementation passes to the MCP server process when it starts a stdio MCP server. | Implementation | Environment variables passed to the MCP server process | Handling of the env field in the configuration | Source | |---|---|---|---| | The official MCP Python SDK | By default, only a list of environment variables that are safe to inherit. On POSIX, these are six: HOME, LOGNAME, PATH, SHELL, TERM, and USER | Passed layered on top of the six | Source code of the stdio client | | The official MCP TypeScript SDK | On POSIX, the same six as the Python SDK | Passed layered on top of the list | Source code of the stdio client | | Qwen Code version 0.21.13 | All of the parent process's environment variables, excluding only three environment variables used to authenticate this product's daemon and its child processes | Passed layered on top of all of them | Source code of the MCP client | | OpenCode version 1.18.18 | The whole of the parent process's environment variables | The environment field is passed layered on top of the whole | Source code of the MCP handling | | OpenHands CLI version 1.16.0 | Reading the dependency path, the six defaults of the official MCP Python SDK | The env field of the configuration is passed to the startup arguments of the official MCP Python SDK | Dependency definition, MCP handling in openhands-sdk, stdio handling in fastmcp | Notes on each row are as follows. - The same default of the official MCP Python SDK is also in the source code of the stdio client in the 1.x series. The latest release is v2.2.0, published on 2026-09-07 - Qwen Code has a file with a function that removes values that are not to be passed (sanitize-child-env.ts). The comments in this file give the reason for limiting what is excluded to internal values. The reason is that commands users run in the shell depend on inheriting third-party credentials - Qwen Code 0.24.6 was published on 2026-09-26. On the main branch after that as well, the Qwen Code CLI is still built to inherit the whole of the parent process's environment variables - On OpenCode's dev branch as well, the design is the same. The latest release is v1.18.33, published on 2026-09-28 - OpenHands CLI version 1.16.0 pins openhands-sdk to 1.21.0 in its dependency definition. The MCP handling in openhands-sdk passes the configuration to fastmcp's MCP configuration type. The current source code of fastmcp passes the env field of a stdio MCP server's configuration as it is to the startup arguments. The startup arguments are passed to the official MCP Python SDK. The latest release of OpenHands CLI is 1.16.0, published on 2026-05-08, the same version used in the measurements The OpenHands CLI row is an explanation obtained by reading the source code of the MCP part. This explanation fits best with the observation that the environment variables did not arrive when the parent was OpenHands CLI. However, the exact versions of fastmcp and the official MCP Python SDK installed at the time of the measurements have not been checked. So this explanation is not the result of checking the whole route in the measurement environment. The observed fact is limited to the way these measurements were started. When the parent was OpenHands CLI version 1.16.0, the parent process's environment variables did not reach the MCP server. The official documentation of the three coding agents describes the field for passing environment variables. The pages checked are the following three. - Official documentation on MCP servers for OpenHands CLI - Official documentation on MCP servers for Qwen Code - Official documentation on MCP servers for OpenCode On the other hand, none of the three pages says whether a started MCP server receives the whole of the parent's environment. The inheritance behavior is determined only by the implementation in the source code, and it can change between versions. Both designs have reasons behind them. The default in the official SDKs narrows the inherited environment variables to avoid leaking secret values. Qwen Code limits what it excludes to values for the product's internal use. The comments in the source code give the reason that commands users run in the shell depend on inheriting third-party credentials. Qwen Code version 0.21.13 passes the parent's environment variables only through that same function when it starts an MCP server as well. The measurement script after the fix does not depend on either design. This is because the measurement script writes the API key in the env field of the MCP server configuration and passes it explicitly. The fix that passes the API key explicitly in the env field of the MCP server configuration, and the re-measurement This section shows the fix to how the API key is passed and the results of the re-measurement after the fix. The AI agent fixed the measurement script. After the fix, the measurement script writes the API key that the child agents use in the env field of the MCP server configuration and passes it explicitly. The AI agent also fixed, in the same way, a separate measurement script that explicitly makes the parent agent call the parallel-execution MCP server. That values written in the env field reach the MCP server in all three coding agents had already been confirmed earlier, by passing a different value. The AI agent set up a unit test to confirm that the API key is written in OpenHands CLI's mcp.json. After that, it extended the unit test to all three coding agents. The following examples show the form for writing the API key in the env field of the MCP server configuration, written from scratch to match the form in the official documentation. The server name, the environment variable name, and the value are placeholders. The first is the Qwen Code form, and OpenHands CLI's mcp.json uses the same form. The second is the OpenCode form. { "mcpServers": { "parallel-runner": { "command": "python", "args": ["-m", "example_parallel_runner"], "env": { "EXAMPLE_API_KEY": "replace-with-your-api-key" } } } } { "$schema": "https://opencode.ai/config.json", "mcp": { "parallel-runner": { "type": "local", "command": ["python", "-m", "example_parallel_runner"], "environment": { "EXAMPLE_API_KEY": "replace-with-your-api-key" } } } } After the fix, the AI agent measured two conditions with OpenHands CLI as the parent again, once each. The two conditions differ in whether the set of instruction files for agents is in place. The set of instruction files for agents is the group of AGENTS.md, skill files, and memory files placed in the working directory. The following table shows the results of the re-measurement under the two conditions. | Condition | Child agents launched | Child agents completed | Tasks passed | Time taken | |---|---|---|---|---| | Set of instruction files in place | 8 | 8 | 8 of 8 | 556.1 seconds | | Set of instruction files not in place | 16 in total | 1 in total | 8 of 8 | 761.1 seconds | The time limit was 900 seconds. Under the condition with the set of instruction files in place, the parent agent called the parallel-execution MCP server on its own. All eight child agents completed. This was the first run, under a condition with OpenHands CLI as the parent, in which the parent agent called the parallel-execution MCP server on its own and every child agent launched completed. Under the condition without the set of instruction files, a call also occurred, and there were three run record files. Of the 16 child agents launched under this condition, 15 did not complete, for reasons other than the API key. The research records say that these 15 include the pattern of fixing an error in the orchestration script and running it again. The research records also say that observation of the causes of these 15 will continue. From these results, the AI agent concluded as follows. The cause of 0 completed child agents was not a difference in capability between the coding agents but the way the measurement script was built. In response to feedback from the reviewing AI agent, the AI agent rewrote the scope of the conclusion. The reviewing AI agent is an AI agent that is started after a PR is created, separately from the agent that did the implementation. It examines the diff read-only. In the rewritten conclusion, the AI agent stated definitively that the problem of every child agent failing because of the API key had been resolved. It separated out the remaining 15 incomplete child agents as having reasons other than the API key. The two re-measurement runs were recorded separately from the comparison totals, as additional measurements for isolating the cause. The re-measurement was one run for each condition. So these records cannot tell how often every child agent launched completes when the parent is OpenHands CLI. A mechanism that does not hide failures, and unit tests that prevent the same omission This section explains the mechanism that does not hide failures and the unit tests that prevent the same omission, both added in PRs after the API key fix. The AI agent added handling for when the API key is needed but missing. In this case, the parallel-execution MCP server stops before launching child agents and marks the run record file as failed. It does not treat the stop as the failure of one child agent, and marks the whole call as failed. The PR that added this handling went through three reviews. 1. In the first review, the reviewing AI agent demonstrated a problem by actually running the code. The problem was that the exception from the stopping logic was treated as the failure of one child agent, and the status in the run record file was left as normal 2. In the second review, the reviewing AI agent showed that with the parallel and sequential ways of writing the orchestration, the same exception was still treated as normal 3. In response to the third review, the AI agent changed the stopping logic so that, once it runs, it leaves a failure flag in the run record file. This flag remains even if the orchestration script catches the exception and swallows it In the same PR, the AI agent removed the logic that, when the API key was empty, overwrote the caller's environment variable with an empty string. After the fix, when there is no API key, the parallel-execution MCP server passes the authentication in the caller's environment to the child agents as it is. In the same PR, the AI agent also added logic that automatically puts a flag on the rows the measurement script records, indicating that there was a defect on the measurement side. Rows with this flag are not counted in the totals as observations of an agent's capability. The flag is added when a task run meets all three of the following conditions. - The exit code is not 0 - The tests were run 0 times - The total time from start to scoring is less than 15 seconds The conditions were made narrow because, if failures caused by a lack of capability were excluded from the totals as defects on the measurement side, the totals would look better than the actual capability. The logic that adds this flag is separate from the logic that judges whether the parallel-execution MCP server was called. The same kind of omission also happened with another environment variable. An isolated directory is passed through an environment variable to Qwen Code, the child agent. This directory is used in place of the user's home directory. The measurement script set this value at the point of starting the parent only when the parent was Qwen Code. So this value reached the child agents through environment variable inheritance only when the parent was Qwen Code. When the parent was one of the other two, the value did not exist at the point of starting the parent, so it did not reach the child agents. The AI agent changed this value as well so that it is written in the env field of the MCP server configuration and passed explicitly for all three coding agents. It confirmed this setup with a unit test. The reviewing AI agent pointed out two places with the same omission. The two places were a separate measurement script and an integration test. The AI agent fixed both places in the same way. In a later PR, the AI agent consolidated the logic that builds the env field of the MCP server configuration into a single function. This is because, if callers set values separately, the same omission happens when the number of places to register grows. The unit test confirms that the set of environment variable names in the env field exactly matches a defined set. If a new credential is written into a configuration file without anyone noticing, this unit test fails. Protection added because the API key is now written in the env field This section explains the new risk created by the fix that passes the API key explicitly, and the protection added in the same PR. After the fix, the API key written in the env field is written as a raw value into a configuration file inside the working directory. The reviewing AI agent returned seven findings on this PR. Each finding is given one of four severity levels: Critical, High, Medium, or Low. All seven had Medium severity. In response to the findings among the seven that concerned protecting the API key, the AI agent fixed the following two things. - It closed off the routes by which the API key value could get into the measurement script's records and into commits. It added logic that mechanically masks values that look like API keys before the records are written, and confirmed this with a unit test. It also fixed the check that runs before a commit and looks for API key values mixed in, staged a file containing an API key value, and confirmed that the commit was rejected - It changed the setup so that the API key is written only in the configuration of the parallel-execution MCP server, which launches the child agents. The key is not written in the configuration of other MCP servers that do not use it. This setup was also confirmed with a unit test Two rules for designing comparison measurements This section presents two rules for designing comparison measurements that can be generalized from these records. The research records contain no statement in this form. 1. Before starting comparison measurements, identify the implicit behaviors that differ between coding agents. Whether the parent process's environment variables reach the MCP server is one of those behaviors 2. Do not leave credentials to environment variable inheritance, and pass them explicitly from the measurement script The basis for rule 1 is the observation that only when the parent was OpenHands CLI did the environment variables fail to arrive, and the source code in which the inherited range differs by implementation. The basis for rule 2 is the facts written in the section on the fix that passes the key explicitly in the env field and the re-measurement, and in the section on the mechanism that does not hide failures. As a supplementary note, the same kind of failure also happened on an external evaluation execution platform. The AI agent ran Qwen Code over ACP on the external evaluation execution platform. ACP is a public protocol for connecting editors and coding agents. This run got as far as starting Qwen Code and connecting over ACP, and then failed with an error at the authentication stage. The task score was 0. In the logic that started the execution platform at that time, the AI agent could not find a configuration item for passing the model's authentication information to the container. The research records treat this as the cause. The measurement script in the same repository had gotten through the same authentication by passing the environment variables and a copy of the configuration file. The AI agent recorded this run as a defect on the measurement side, not as an observation of capability. The inheritance behavior of the parent process's environment variables can be checked in the source code of the version in use, not in the official documentation. The part to read is the logic that starts a stdio MCP server. It is a good idea to check whether that logic passes the whole of the parent's environment or narrows it with a list, and only then use that version of the coding agent in comparison measurements. Materials I referred to The public materials referred to in the body are as follows. All of them were checked on 2026-09-29. - The source code of the stdio client in the official MCP Python SDK github.com/modelcontextprotocol/python-sdk/blob/m… - The source code of the stdio client in the 1.x series of the official MCP Python SDK github.com/modelcontextprotocol/python-sdk/blob/v… - The release page of v2.2.0 of the official MCP Python SDK github.com/modelcontextprotocol/python-sdk/releas… - The source code of the stdio client in the official MCP TypeScript SDK github.com/modelcontextprotocol/typescript-sdk/bl… - The source code of the MCP client in Qwen Code version 0.21.13 github.com/QwenLM/qwen-code/blob/v0.21.13/package… - The source code of the function that removes values not to be passed, in Qwen Code version 0.21.13 github.com/QwenLM/qwen-code/blob/v0.21.13/package… - The source code that checks authentication in non-interactive mode, in Qwen Code version 0.21.13 github.com/QwenLM/qwen-code/blob/v0.21.13/package… - The release page of Qwen Code 0.24.6 github.com/QwenLM/qwen-code/releases/tag/v0.24.6 - The source code of the MCP handling in OpenCode version 1.18.18 github.com/anomalyco/opencode/blob/v1.18.18/packa… - The release page of OpenCode v1.18.33 github.com/anomalyco/opencode/releases/tag/v1.18.… - The dependency definition of OpenHands CLI version 1.16.0 github.com/OpenHands/OpenHands-CLI/blob/1.16.0/py… - The release page of OpenHands CLI 1.16.0 github.com/OpenHands/OpenHands-CLI/releases/tag/1… - The source code of the MCP handling in openhands-sdk version 1.21.0 github.com/OpenHands/software-agent-sdk/blob/v1.2… - The source code of the stdio handling in fastmcp github.com/PrefectHQ/fastmcp/blob/main/fastmcp_sl… - The official documentation on MCP servers for OpenHands CLI docs.openhands.dev - The official documentation on MCP servers for Qwen Code qwenlm.github.io - The official documentation on MCP servers for OpenCode opencode.ai --- Originally published at The Future of Humans, AI, and the Web, a site where my research and development is recorded and analyzed by a human and an AI.

FastMCP vira função Python em tool MCP com um decorator. O schema sai dos type hints. pip install fastmcp, @mcp.tool, teste com Client(mcp). Rodei a v4.0.11 num mascarador de CPF sintético: schema gerado, chamada ok. Minha função era um regex. Não detecta PII. PrefectHQ/fastmcp