This AI Turns URLs into Viral Videos

Cole Medin recently showcased a groundbreaking system live, demonstrating the seamless conversion of a single URL into a fully rendered, production-grade video. Imagine feeding a Hacker News story, an intricate blog post, or a detailed product page into a machine and receiving a complete MP4 in return. This isn't theoretical; Medin's workflow autonomously fetches source content, meticulously plans scenes, crafts dynamic voiceovers, music, and sound effects, and even writes the underlying TypeScript composition.

Traditional video production remains a laborious, costly, and time-intensive endeavor. It demands a dedicated team of scriptwriters, videographers, editors, and sound engineers, often stretching timelines and budgets for even short-form content. Medin's innovation radically redefines this paradigm, automating every facet from initial concept ingestion to final render, bypassing the conventional bottlenecks of manual labor and specialized skill sets.

This automated pipeline offers a compelling glimpse into the future of digital content creation. The system’s speed and efficiency suggest a world where the only limiting factor is the ideation itself, not the laborious execution. Content creators can now focus solely on narrative and strategy, confident that the technical heavy lifting of video production is handled instantaneously, enabling unprecedented scale and responsiveness.

Achieving this level of automation requires a powerful synergy of advanced technologies. Medin’s 14-step workflow, orchestrated within Archon (archon.diy), leverages Claude Code for intelligent planning, building, and quality assurance. This sophisticated AI agent guides the entire process, from content analysis to auto-fixing critical issues. For the visual and auditory output, the system integrates Cartesia and ElevenLabs for voice, music, and sound effects, alongside Remotion for crafting the TypeScript video compositions.

Remotion, a programmatic video framework, is crucial here, transforming AI-generated instructions into polished video. A 39-rule best-practices skill injected into the build node ensures the generated compositions feature real transitions, dynamic durations, and proper hooks, elevating the output beyond typical "AI slop." This meticulous approach guarantees high-quality, professional-looking videos directly from a simple URL.

Cole Medin’s viral video generator is no simple AI tool; it represents a sophisticated orchestration of cutting-edge platforms. At its core, a powerful trinity drives the entire operation: Archon, Claude Code, and Remotion. This isn't a single AI creating magic, but a meticulously designed, agentic workflow that transforms a URL into a polished video, setting it apart from simpler generative systems.

Archon serves as the workflow’s central nervous system, acting as the "project manager" for the entire operation. Medin’s open-source workflow engine defines and executes the complex, 14-step development process. It reliably runs the automated sequence, allowing for parallel execution across isolated branches and offering a web dashboard to monitor every agent activity.

Claude Code steps in as the "artist and developer," an AI coding assistant with a deep understanding of codebases. It performs the critical creative and technical tasks: planning video scenes, generating voiceovers, music, and sound effects. Crucially, Claude Code also writes the TypeScript composition for the video, then conducts a quality pass, auto-fixing critical issues before rendering.

Remotion functions as the "canvas," the programmatic video framework that brings the visual elements to life. Built on React, Remotion takes the TypeScript compositions generated by Claude Code and renders them frame by frame into high-quality MP4 files. This allows the system to leverage web development paradigms for dynamic durations, real transitions, and proper hooks, avoiding generic "AI slop."

The true differentiator lies in this seamless integration and orchestrated synergy. Archon manages Claude Code, which in turn feeds Remotion, creating an end-to-end production pipeline previously requiring a team of human specialists. This sophisticated workflow ensures the output videos are not just generated, but thoughtfully composed and refined.

Beyond the core trio, supporting AI services enhance the output. Cartesia and ElevenLabs handle the nuanced audio production, generating realistic voiceovers, bespoke music tracks, and crucial sound effects. These specialized tools integrate directly into Claude Code’s generative process, completing the immersive video experience.

Cole Medin’s demonstration unveils a robust 14-step workflow, meticulously engineered to transform any URL into a polished MP4 video. This intricate process unfolds across four distinct phases: Ingestion, Planning, Generation, and Production, meticulously orchestrated to ensure high-quality output. Each phase leverages specialized AI capabilities and development frameworks to automate complex video creation.

The journey begins with Ingestion, where the system fetches source content directly from the provided URL, whether it’s a Hacker News story, a detailed article, or a product page. Following this, the Planning phase kicks in, where Claude Code intelligently dissects the fetched content. It outlines the narrative, plans individual scenes, and structures the video’s flow, acting as the project’s initial creative director.

Moving into the Generation phase, the system synthesizes all necessary media assets. This involves generating a compelling voiceover, selecting appropriate background music, and adding relevant sound effects, primarily utilizing Cartesia and ElevenLabs. Simultaneously, the system writes the TypeScript composition for Remotion, adhering to a sophisticated "39-rule best-practices skill" injected into the build node. This ensures the generated compositions feature authentic transitions, dynamic durations, and proper hooks, elevating the video quality far beyond typical AI-generated content.

The Production phase culminates the process with crucial quality control. The system performs an autonomous quality pass, meticulously identifying and then auto-fixing critical issues, showcasing its impressive agentic nature. This self-correction mechanism, powered by Claude, prevents common AI "slop" and ensures a professional finish before rendering the final MP4. This proactive QA loop guarantees reliability and minimizes manual intervention.

Archon stands as the workflow’s backbone, orchestrating every single node within this 14-step pipeline. As an open-source workflow engine for AI coding agents, Archon ensures a reliable and repeatable process from start to finish. It manages the handoffs between different AI models and frameworks, guaranteeing seamless execution and robust output, much like an n8n for code.

For deeper insights into the underlying AI, one can explore Claude AI by Anthropic, which serves as a core intelligence for planning and quality assurance in this system. This end-to-end automation exemplifies a production-grade AI workflow, transforming raw web content into engaging video with unprecedented efficiency and quality.

Remotion introduces a fundamental paradigm shift, reframing video creation not as a graphical design task but as a software development problem. Developers now write TypeScript compositions to define every visual element, transition, and timing, effectively coding their videos from the ground up. This turns dynamic media into a programmable asset.

This programmatic approach brings immediate, transformative benefits. Teams can implement robust version control using Git, tracking every change, reverting to previous states, and collaborating seamlessly on video projects just as they would with any codebase. Scalability dramatically improves; rendering video compositions on servers allows for mass production and on-demand generation, bypassing the bottlenecks of local machines.

Furthermore, video elements become reusable React components. This modularity means developers build libraries of intros, outros, lower-thirds, and complex animations, accelerating future projects and ensuring brand consistency across hundreds or thousands of videos. Automation becomes a reality through CI/CD pipelines, enabling continuous video generation and updates tied directly to data changes or content feeds.

Remotion’s reliance on React is a strategic advantage, tapping into an enormous existing ecosystem. Developers can leverage their familiar React skills, tools, and libraries to build sophisticated video experiences. This access to a vast community and established development practices significantly lowers the barrier to entry for programmatic video.

This method stands in stark contrast to the limitations of timeline-based GUI editors. Traditional editors require manual, frame-by-frame adjustments, making large-scale automation or data-driven content generation impractical. By treating video as code, Remotion unlocks dynamic, personalized content at a scale previously unimaginable, pushing the boundaries of what automated media creation can achieve.

Claude Code functions as the workflow’s AI film director, an agentic intelligence orchestrating the entire video generation process from a simple URL. It moves beyond basic content summarization, actively understanding, planning, and executing complex creative tasks within the Archon framework. This sophisticated AI acts as the central brain, guiding the transformation of text into dynamic visual narratives.

Initially, Claude Code ingests the source material from the provided URL, whether a Hacker News article, product page, or detailed blog post. Its first critical task involves deep comprehension of the text, dissecting the content to identify core themes, extract salient information, and structure a compelling narrative arc suitable for video. This involves not just summarization, but strategic scene planning and storyboarding.

With a narrative blueprint in hand, Claude then writes Remotion TypeScript composition. This demands more than just code generation; it necessitates embedding proper hooks for dynamic content insertion, implementing precise timing for scene transitions, and ensuring overall flow aligns with professional video production standards. Claude dictates every visual element and its temporal relationship, effectively scripting the video frame by frame.

A crucial element enhancing Claude’s output is the 39-rule best-practices 'skill' injected during the build phase. This comprehensive set of guidelines prevents the generation of generic, visually uninspired "AI slop." Instead, Claude leverages these rules to create sophisticated compositions featuring: - Real transitions - Dynamic durations - Proper content hooks

These rules empower Claude to craft high-quality, non-generic video sequences, elevating the aesthetic and functional quality of the final product.

Claude’s agentic capabilities extend to a crucial quality assurance (QA) pass on its own generated code. It identifies critical issues, debugs errors, and autonomously implements fixes, ensuring Remotion composition is robust and render-ready. This self-correction loop is a game-changer for reliable automation, drastically reducing the need for human oversight in debugging code.

This iterative process of generation, evaluation, and self-correction makes Claude Code an indispensable component of the 14-step workflow. It transforms raw web content into polished, dynamically timed video narratives, demonstrating a profound leap in AI-driven creative automation. Claude’s ability to act as a complete "film director"—from story conceptualization to final code QA—underscores its pivotal role in Medin's innovative system.

Archon emerges as the unsung hero orchestrating Medin's complex, 14-step video generation pipeline. This open-source workflow engine transforms a chaotic series of AI agent interactions into a manageable, deterministic process. It ensures each stage, from content ingestion to final MP4 rendering, executes reliably and predictably.

Consider Archon the workflow engine for AI coding agents, akin to "n8n but for code." Instead of connecting APIs or bash scripts, Archon sequences and manages autonomous AI agents, like Claude Code, through multi-step development tasks. This allows for sophisticated automation that goes far beyond simple linear scripts.

Developers define these intricate, multi-step processes using declarative YAML files. This approach allows for easy modification, versioning, and sharing of entire workflows. Teams can iterate on complex AI-driven pipelines with the same rigor applied to traditional software development.

Archon supports robust features critical for production environments. It enables parallel execution across isolated branches, significantly accelerating the overall workflow by running concurrent tasks. This capability is vital for processing multiple video requests or optimizing complex sub-tasks.

A dedicated web dashboard provides real-time monitoring of agent activity, offering granular insights into each step's progress and status. This visibility is indispensable for debugging, performance optimization, and ensuring the system's overall health. Medin’s demonstration highlights Archon’s readiness for demanding, end-to-end AI applications.

Archon’s architecture ensures that AI agents can reliably plan, implement, validate, and review code, even creating pull requests automatically. This level of automation underscores its potential to revolutionize development operations. For example, while Archon handles workflow orchestration, other specialized AI systems like Cartesia AI: Products for real-time, multimodal intelligence. could provide real-time, multimodal intelligence for content analysis or voice synthesis within a broader ecosystem.

This robust orchestration layer makes the entire system resilient and scalable. Without Archon, managing the interplay between content fetching, scene planning, voiceover generation, TypeScript composition with Remotion, and quality assurance would become a monumental, error-prone task. It truly acts as the agentic conductor, ensuring every component plays its part in harmony.

The early days of generative AI produced a flood of content often dismissed as "AI slop"—generic, repetitive, and devoid of professional polish. This common pitfall, characterized by bland visuals and predictable structures, threatens to undermine the utility of AI in creative fields. Cole Medin’s URL-to-MP4 workflow directly confronts this challenge, ensuring its output rises far above the average.

Central to this distinction is a meticulously crafted 39-rule best-practices skill injected directly into Claude Code. This isn't merely a prompt; it's a comprehensive guide that imbues the AI agent with a deep understanding of video production principles. The system leverages these rules during the generation phase, transforming raw content into sophisticated compositions.

These injected guidelines cover critical aspects of professional video creation. They mandate the use of real transitions between scenes, preventing abrupt cuts and enhancing visual flow. The rules also dictate dynamic durations for video segments, ensuring content length adapts intelligently to the underlying information rather than adhering to rigid, arbitrary timings.

Furthermore, the skill set enforces proper application of React hooks within Remotion compositions, guaranteeing robust and efficient code. Aesthetic principles, such as consistent branding elements and optimal text placement, are also integrated, elevating the visual appeal. This proactive "harness engineering" transforms a powerful AI into a highly skilled, albeit automated, video editor.

This strategic injection of expert knowledge is the true differentiator. It enables Claude Code to produce professional-grade video, sidestepping the superficiality often associated with AI-generated media. By codifying design and production best practices, Medin's system proves that directed AI, rather than unrestrained generation, holds the key to high-quality, scalable content creation.

Cole Medin’s live demonstration of his URL-to-MP4 system unveils a profound paradigm shift for the creator economy and modern marketing. This sophisticated workflow, leveraging Claude Code, Remotion, and Archon, fundamentally redefines who can produce high-quality video content. Developers, traditionally outside the realm of video production, are now empowered to become prolific content creators at an unprecedented scale, without needing traditional video editing skills or specialized software.

This technological leap unlocks entirely new categories of dynamic media. Imagine hyper-personalized video advertisements, custom-generated for individual users based on their browsing history, purchase intent, or demographic data, delivering unparalleled relevance. Envision automated news summaries that transform complex articles or live data feeds into engaging video briefings, complete with voiceovers from ElevenLabs and music from Cartesia, all at the push of a button. Consider dynamic product demonstrations, automatically updated and rendered from evolving documentation or product specifications, ensuring every video reflects the latest features without manual intervention.

The core innovation lies in treating video creation as a software development problem rather than a manual artistic endeavor. Developers transition from the laborious task of frame-by-frame editing to designing sophisticated, automated creative systems. They architect the programmatic pipelines, define the TypeScript composition rules for Remotion, and instruct Claude Code on the narrative flow, scene planning, and quality assurance. This approach allows for version control, modularity, and rapid iteration, mirroring best practices in software engineering.

This shift fundamentally reconfigures the creative workflow, moving from bespoke, manual efforts to scalable, code-driven automation. Marketers can now deploy A/B tested video campaigns with unprecedented speed, while content agencies can generate vast libraries of tailored content efficiently. The system promises unparalleled efficiency and consistency, positioning developers not just as builders of software, but as the architects of the next generation of creative output, where content scales with the ingenuity of code.

Cole Medin’s URL-to-MP4 workflow sharply contrasts with off-the-shelf AI video SaaS platforms like InVideo or Synthesia. His custom system, integrating Claude Code, Remotion, and Archon, offers unparalleled control, treating video generation as a deep software development problem.

This custom rig empowers developers to architect every facet of production. Users gain complete command over scene planning, media generation, and TypeScript composition, ensuring videos align perfectly with brand guidelines. Post-setup, the system operates without recurring per-video costs, enabling scalable, cost-effective high-volume output.

Such power demands significant development expertise. Implementing a multi-agent workflow like Medin's requires proficiency in coding, agent orchestration, and debugging. Initial setup and resource investment are substantial, making it

Inspired by Medin's live demo, you can begin constructing your own programmatic video pipeline today. Dive into the open-source blueprint for URL-to-MP4 automation, available on Cole Medin's GitHub repository. This provides a tangible starting point for understanding the intricate 14-step workflow that transforms a simple link into a polished, production-grade video.

Access the core technologies that power this revolution. Explore the official documentation and vibrant developer communities for: - Remotion: The React-based framework that treats video as code, enabling unparalleled precision and scalability. - Archon: The agentic workflow engine orchestrating complex AI tasks with deterministic reliability. - Claude: Anthropic's powerful AI model, serving as the intelligent director for scene planning, script generation, and quality assurance.

As a practical first project, select one of your own blog posts or a favorite article. Challenge yourself to automate its transformation into a concise, branded video summary using Medin's architectural principles. This hands-on experience illuminates the profound power of defining video logic programmatically, moving beyond the limitations of manual editing.

Embrace the paradigm shift from traditional NLEs to code-driven content creation. The synergy between Remotion's declarative video capabilities, Archon's robust orchestration, and Claude's agentic intelligence unlocks unprecedented scale and consistency for media production. Experiment, iterate, and discover how treating video as a software development problem empowers you to build a dynamic, automated media factory. The future of video production is programmatic, and the tools are now at your fingertips, ready for your innovation.

What is the core idea behind this AI video generation workflow?

The core idea is to fully automate the video creation process from a single URL input. It uses an orchestrated system of AI agents and programmatic tools to handle everything from content analysis and scene planning to code generation and final rendering, producing a finished MP4 file.

How does Remotion create videos with code?

Remotion is a framework that allows you to create videos programmatically using React. You build video scenes as React components, and Remotion renders these components frame by frame into a video file, enabling version control, scalability, and automation for video production.

What role does an AI agent like Claude Code play in this process?

Claude Code acts as the 'AI director' and 'developer'. It analyzes the source content, plans the video scenes, generates the voiceover script, and writes the actual TypeScript code for the Remotion composition. It also performs quality assurance and can even auto-fix bugs in its own code.

Is this workflow accessible to non-developers?

This specific workflow is developer-centric, as it involves TypeScript, React, and YAML configurations. However, it represents a paradigm shift where the underlying complexity could eventually be abstracted away, making powerful, customized AI video generation more accessible to a wider audience.