The automation backbone. Two hundred thousand stars of proof that it runs real companies.
First stepSelf host with Docker and build one workflow that runs daily.
The Web Network
One hundred four open source tools for people who build with AI. Pulled from three thousand one hundred eighty eight candidates. Every entry checked against GitHub on October 23, 2026.
Built for the Outsiders. People who would rather own the machine than rent it.
See it. Believe it. Obtain it.
Priced at the typical list price of the paid tool each one replaces. The list itself costs nothing. Typical list prices, October 2026, shown as estimates and counted tool by tool.
104 tools, 7 stack recipes, the Short List of 10, the Prompt Pack, and the AI Edition file your assistant reads.
$1,098 a month in paid tools, counted tool by tool at typical list prices.
$0. Open source, all of it. Every card shows its own saving on its Saves line.
Send the list to your AI. It interviews you, picks your stack, and gives you the first step for each tool.
This page is written for people. The file is written for machines. Download the AI Edition and attach it to Claude or ChatGPT with the prompt. Your assistant interviews you, picks a stack from this list, and gives you the first step for each tool. It works only from this list. Nothing invented. The file carries the same numbers as this page: $1,098 a month in paid tools replaced, tool by tool, against $0 to run the list.
Works with every AI. Paste it, upload it, or point your agent at this page. The page itself carries the full text for agents that browse.
Published by The Web Network · thewebnetwork.app
Five prompts that turn the file into a working session. Copy one, paste it with the file attached, and answer what your AI asks.
Read the attached Web Network repo list. Ask me what I am trying to do, then recommend one stack from its Stack Recipes, or compose one from the catalog. For each tool, tell me what it replaces and give me the exact first step. Use only repos from the file.
Read the attached list first. Then interview me. Five questions about what I make, where I publish, and what I pay for today. Then choose my stack and defend the choice.
Take the stack from the attached list. Walk me through it one tool at a time. Give me the first step only. Check my result before you give me the next tool.
From the attached list, compare the tools in one category for my use case. Name one winner. Say why in three lines. Use only tools from the file.
Build me a thirty day plan from the attached list. One new tool per week. Daily actions under one hour. Every week ends with something published.
Find the job. The list is organized by the work, not the technology. Start at the job you are trying to get done.
Read the lines. Every tool states what it does, what it replaces, what it saves, and how an agent uses it. The paid tool it replaces is named and priced, so the saving is concrete.
Hand it to your AI. The full list exists as one file written for an AI to read. Give the file to your assistant and it builds the stack with you.
If you make video: the Short List, then the AI Editor Stack, then the Faceless Channel Stack.
If you market: the Growth Stack, then Research and content intel.
If you build product: the Agent Starter Stack, then For the builders.
If you sell services: the Automation Stack, then the Growth Stack.
Ten tools. If a Climber installs nothing else from this page, install these. Each one replaces a paid subscription or a paid hire, and each one is proven at scale.
The automation backbone. Two hundred thousand stars of proof that it runs real companies.
First stepSelf host with Docker and build one workflow that runs daily.
The standard every other skill is measured against. Start here and the rest of the list makes sense.
First stepAdd a skill folder to .claude/skills in your project. Claude Code loads it automatically.
The transcription standard. Every caption and clipping workflow starts here.
First steppip install openai-whisper, then transcribe one file.
Whisper at production speed. The version you deploy.
First steppip install faster-whisper and transcribe the same file. Compare the clock.
The ElevenLabs replacement that runs on your machine. Cloning, design and dubbing in one place.
First stepDownload the app and clone your voice from a short sample.
A production system, not a demo. Twelve pipelines that finish videos.
First stepClone the repo and run the quickstart pipeline on a test topic.
The standard for video in code. Every serious programmatic video stack touches it.
First stepRun npx create-video, then render the starter composition.
A marketing department as a skill pack. Copy, SEO, analytics and growth in one install.
First stepAdd the skills to Claude Code and run a CRO pass on your main page.
What the internet actually said in the last thirty days, synthesized in minutes. Research stops being a day of tabs.
First stepInstall the skill and research your next topic before you write a word.
Reverse engineering at agent speed. The repo of the year for a reason.
First stepPoint it at one app you rely on and read what it finds.
Tools compound when they are wired together. These are seven working combinations, one per outcome. Pick the outcome. Build the stack.
Wired together, the seven stacks replace about $700 a month in subscriptions at typical list prices. Over $8,000 a year. Prices checked October 2026 and labeled as typical. The tools cost nothing.
Get an AI that actually knows how to work, in one sitting.
Replaces a paid course and a month of weekends.
Edit reels and Shorts by talking to your AI instead of touching a timeline.
Replaces about $50 a month in subscriptions.
Research → script → voice → video → clips, end to end.
Replaces about $120 a month in subscriptions.
A full ElevenLabs replacement that runs on your own machine, free.
Replaces about $99 a month in subscriptions.
AI video and pages that don't look AI-made.
Replaces a designer day rate on every project.
Marketing, SEO, ads and posting, run by skills, not retainers.
Replaces about $250 a month in subscriptions.
Wire it all together so it runs without you.
Replaces about $100 a month in subscriptions.
Skills are instruction packs. Install one and your AI works to that standard every time.
The biggest skills framework out there. A full working methodology for agents.
Plugs into your AI View on GitHubBattle-tested skills from a top educator's own setup. Works across Claude, Codex, Cursor.
Plugs into your AI View on GitHubAgent harness optimization system: skills, instincts, memory and security tuning for Claude Code, Codex and Cursor.
Plugs into your AI View on GitHubA single CLAUDE.md file that improves Claude Code behavior, built from Andrej Karpathy observations on how agents fail.
View on GitHubThe official Anthropic skills repo. The standard everyone else builds on.
Plugs into your AI View on GitHubClaude Code itself: the agentic coding tool that lives in your terminal.
Plugs into your AI View on GitHubDesktop all in one assistant for Claude Code, Codex, OpenCode and other harnesses: providers, skills and sessions in one place.
Plugs into your AI View on GitHubCodex CLI itself: OpenAI lightweight coding agent for the terminal.
Plugs into your AI View on GitHubMakes your agent answer in compressed caveman-speak and cuts token usage ~65%. Funny, and it actually saves money.
Plugs into your AI View on GitHubProduction grade engineering skills for AI coding agents, from a Chrome engineering lead.
Plugs into your AI View on GitHubPersistent memory across sessions for agents: captures what your agent did and recalls it in the next session.
View on GitHubDirectory of MCP servers, how you plug tools (scrapers, socials, databases) into your agent.
Plugs into your AI View on GitHubHand-picked best Claude Code resources. The list to mine.
View on GitHub1,000+ skills in one curated directory when you want to browse.
Plugs into your AI View on GitHubThe official Codex skills catalog, from OpenAI.
Plugs into your AI View on GitHubProduction systems an agent can run end to end.
One click short video generation: give it a topic or keyword, it writes, voices, captions and assembles an HD short.
View on GitHubOpen-source agentic video production system: 12 pipelines, 100+ tools. The closest thing to an AI production studio.
View on GitHubWrite HTML, render real video. Built for agents, from the HeyGen team.
View on GitHubMake videos programmatically with React. The industry standard for code-driven video.
View on GitHubEdit videos by telling a coding agent what you want.
View on GitHubCreator workspace for video translation and dubbing, powered by Codex. Localize content into new markets.
View on GitHubCinematic product videos with Remotion: 152 shot recipes, 209 motion presets.
Plugs into your AI View on GitHubAI narration and re editing for long video: commentary scripts and automatic cuts for film and series content.
View on GitHubOne topic in, finished Vox-style paper-collage explainer/ad video out, end to end.
View on GitHubEdit reels and Shorts with Codex or Claude Code using transcription-based editing. Small, sharp, built for exactly this group.
Plugs into your AI View on GitHubTurns a vague idea into a second-by-second storyboard spec your agent can execute.
Plugs into your AI View on GitHubA skill that teaches Claude to build professional Remotion videos for you.
Plugs into your AI View on GitHubHeyGen's own agent skills for avatar creation and video production.
Plugs into your AI View on GitHubThe unglamorous tools that return hours every week.
The command line audio and video downloader. Pulls media from YouTube and a thousand other sites.
View on GitHubThe open source CapCut alternative: a real timeline editor in the browser, free.
View on GitHubThe Swiss Army knife for fast, lossless video trimming. Instant cuts, no re-encode.
View on GitHubEdit video in Python: cuts, overlays, text, composites.
View on GitHubTranscribe a video, then clip the best moments with AI help.
View on GitHubOne-click batch generation of short marketing/content videos.
View on GitHubAutomatically cuts silences and dead air from footage.
View on GitHubLocal AI video editor with a real timeline, edit by chatting with it.
Plugs into your AI View on GitHubOpen models for image and video. No subscription meter running.
The most powerful modular diffusion GUI and backend: node based workflows for image, video and audio models.
View on GitHubThe library underneath almost every open image/video/audio model.
View on GitHubFree AI image & video generation studio with 600+ models. Start here.
View on GitHubOpen-source AI avatar / digital human toolkit. Clone a presenter, generate videos offline.
View on GitHubText-to-video and image-to-video generation (CogVideoX).
View on GitHubTencent's heavyweight open video generation model.
View on GitHubFast AI video generation for modest GPUs: Wan, LTX, Hunyuan and MiniMax models tuned for consumer hardware.
View on GitHubImage generation model that's unusually good at rendering text in images.
View on GitHubUnlimited-length talking-head video from an image or existing video.
View on GitHubMusic generation foundation model: full songs with vocals and lyrics, in many styles and languages.
View on GitHubA studio front-end for image/video generation models, one prompt bar for all of them.
View on GitHubCloning, dubbing and transcription that run on your own machine.
The transcription standard. Every caption workflow starts here.
View on GitHubFully local ElevenLabs alternative: voice cloning, voice design, video dubbing.
View on GitHubClone a voice from as little as 1 minute of audio.
View on GitHubOpen-source AI voice studio. Clone, dictate, create.
View on GitHubMicrosoft's frontier open voice AI.
View on GitHubInstant voice cloning from MIT/MyShell.
View on GitHubFree, offline speech-to-text app. Dictate anywhere, no cloud.
View on GitHubState-of-the-art open text-to-speech.
View on GitHubResemble AI's open TTS, expressive, production-ready.
View on GitHubWhisper, but much faster. Use this one in production.
View on GitHubVoice cloning TTS across 600+ languages.
View on GitHubTiny, fast TTS model. Great for bulk narration.
View on GitHubThe difference between AI video and a film is motion judgment. These carry it.
Design intelligence skill for pro-level UI/UX output.
Plugs into your AI View on GitHubGives coding agents actual design taste: art direction, references, visual checks.
Plugs into your AI View on GitHub3Blue1Brown's animation engine for explainer videos.
View on GitHubEditorial diagram design skill for coding agents: forty four diagram types in clean HTML and SVG.
Plugs into your AI View on GitHubThe community edition, easier on-ramp to the same style.
View on GitHubModern animation library for React/JS (formerly Framer Motion).
View on GitHubAnti-AI-slop design skill. Your landing pages stop looking generated.
Plugs into your AI View on GitHubThe pro JavaScript animation library behind most award-winning web motion.
View on GitHubAnimate ideas and presentations with TypeScript.
View on GitHubUniversal motion principles (timing, easing, Disney principles) as a skill your agent follows. This is how AI video stops looking like AI video.
Plugs into your AI View on GitHubSkill that turns AI-slop videos into films that look designed by a motion designer.
Plugs into your AI View on GitHubGrowth work as skills and systems, not retainers.
Marketing skills for agents: CRO, copywriting, SEO, analytics, growth. The content-marketing pack.
Plugs into your AI View on GitHubSelf-hosted social media scheduling with agent features. Post everywhere, no per-seat SaaS bill.
View on GitHubModern link attribution, know exactly which post drove the click/sale.
View on GitHubOpen-source Semrush/Ahrefs alternative.
View on GitHubFull SEO skill: technical audits, schema, E-E-A-T, 26 sub-skills.
Plugs into your AI View on GitHubSelf-hosted email marketing server. Own your list, no Mailchimp rent.
View on GitHubOpen-source marketing automation (drips, lead scoring, campaigns).
View on GitHubPaid-media operations skill across 12 ad platforms (Google, Meta, TikTok…).
Plugs into your AI View on GitHubAgent skills for outlier-post hunting, comment mining, competitor teardowns, ad libraries.
Plugs into your AI View on GitHubSelf-hosted agent that runs a YouTube channel end to end: research, script, produce.
View on GitHubYouTube creator skill: channel audits, retention scripts, thumbnails, content strategy.
Plugs into your AI View on GitHubTikTok marketing skills: 3-second hooks, captions, trend mapping. Small but on-target.
Plugs into your AI View on GitHubKnow what is already working before you make anything.
Reverse-engineer anything with agents, from app behavior down to protocols. #1 trending on GitHub, and it earns it.
View on GitHubResearches any topic across Reddit, X, YouTube, HN and the web, then synthesizes it. Content intel gold, know what's already working before you create.
Plugs into your AI View on GitHubDownloads image galleries and collections from social and image platforms, with metadata.
View on GitHubDownloads Instagram posts, stories and metadata for research and archiving your own account.
View on GitHubLets your agent scrape social media, search, maps and e-commerce on demand.
Plugs into your AI View on GitHubWire the tools together so the system runs without you.
The workflow automation platform with native AI. Connect your apps, automate the boring parts.
View on GitHub280+ ready-to-use n8n workflows for Gmail, Telegram, Slack, WhatsApp, Sheets and more.
View on GitHubAI workflow automation with ~400 MCP connectors. The other strong n8n alternative.
View on GitHubThe layer underneath. Backends, frameworks and agent infrastructure.
Run open models on your own machine with one command: Llama, Qwen, DeepSeek, Gemma and more.
View on GitHubMakes your agent think like the laziest senior dev, the best code is code never written.
Plugs into your AI View on GitHubBuild agentic workflows and AI apps on a visual canvas.
View on GitHubA friendly chat interface for your local models and APIs: users, files, RAG and tools in one self hosted app.
View on GitHubGarry Tan's exact Claude Code setup: 23 opinionated tools that act like a whole team.
View on GitHubThe component library everyone builds product UI with now.
View on GitHubAI browser agents in Python. Automate anything with a web page.
View on GitHubDatabase, auth and storage platform when your MVP outgrows a spreadsheet.
View on GitHubA real backend in a single file. Fastest way to ship an MVP.
View on GitHubFramework for orchestrating role playing agent crews that divide work and finish tasks together.
View on GitHubSmall repositories with one sharp use each.
Skill for premium scroll-driven websites that feel expensive.
Plugs into your AI View on GitHubSkill that automates CapCut (JianYing) editing through an agent. Successor to the original repo, which came down in October 2026.
Plugs into your AI View on GitHubControl DaVinci Resolve from your AI agent via MCP. If you edit in Resolve, this is a cheat code.
Plugs into your AI View on GitHubClaude Code skills for DaVinci Resolve workflows. Pairs with the bridge above.
Plugs into your AI View on GitHubThis page is written for people. The file is written for machines. Download the AI Edition and attach it to Claude or ChatGPT with the prompt. Your assistant interviews you, picks a stack from this list, and gives you the first step for each tool. It works only from this list. Nothing invented. The file carries the same numbers as this page: $1,098 a month in paid tools replaced, tool by tool, against $0 to run the list.
Works with every AI. Paste it, upload it, or point your agent at this page. The page itself carries the full text for agents that browse.
Published by The Web Network · thewebnetwork.app