Build-it-yourself guides

Blueprints.

A blueprint is a practical build guide: follow the steps, work through the code, and finish with something you can run.

33 build guides · each ends with something you can run
  1. beginner · 1 stepRun an Open Model with OllamaThree commands. A local LLM running on your laptop, with an OpenAI-compatible HTTP API ready to call from anywhere.
  2. intermediate · 8 stepsBuild Your Own MCP ServerBuild a personal bookmarks server that lets Claude save, search, and organize your links — Node.js, SQLite, and the official MCP SDK.
  3. intermediate · 9 stepsBuild a RAG Pipeline from ScratchBuild a document Q&A system with plain Python, Supabase pgvector, OpenAI embeddings, and Claude — no frameworks, just code you can read end to end.
  4. intermediate · 7 stepsBuild Agentic RAG with Supabase + GeminiUpload text, PDFs, or a URL; ask questions; get answers with sources. A Next.js page on Vercel plus two Supabase Edge Functions — a write phase (chunk, embed with Supabase's built-in model, store in pgvector) and a read phase (retrieve nearest chunks, ground Gemini) — with the model key kept server-side. Deployable in about ten minutes.
  5. intermediate · 9 stepsShip an AI Agent with ClaudeBuild a research agent that searches the web, summarizes articles, and compiles reports — deployed as a REST API.
  6. intermediate · 8 stepsBuild a Code Review BotCreate a GitHub Action that reviews pull requests with Claude — catches bugs and posts inline comments automatically.
  7. advanced · 11 stepsCreate a Multi-Agent SystemBuild a content production pipeline with CrewAI — a researcher, writer, and editor that collaborate to produce articles.
  8. advanced · 10 stepsDeploy Your Own Open-Source LLMRun Llama 3.1 locally with Ollama, benchmark it, then deploy a production API with vLLM and Docker.
  9. intermediate · 8 stepsBuild Your Own AI CLI ToolBuild a terminal assistant called "ask" — answer questions, explain files, generate shell commands, and pipe in data.
  10. intermediate · 9 stepsBuild an Agent with Google ADKBuild a research agent with Google's Agent Development Kit and Gemini — custom tools, sessions, and one-command deploy to Cloud Run.
  11. intermediate · 10 stepsBuild a LangGraph Agent TeamBuild a job application assistant where four AI agents collaborate via LangGraph — analyze postings, tailor resumes, write cover letters, and review everything.
  12. advanced · 9 stepsBuild a Claude Code Harness ExtensionBuild a code-review swarm as a Claude Code plugin — sub-agents, hooks, custom slash commands, an MCP tool server, and a background audit worker, packaged as a publishable plugin.
  13. advanced · 11 stepsBuild a Shared-Skills MCP ServerA team snippet library exposed as a remote MCP server: Supabase OAuth 2.1 with dynamic client registration, workspaces with RLS-enforced sharing, and one `claude mcp add` to wire your whole team's saved prompts into Claude.
  14. beginner · 7 stepsBuild a Team-Token MCP ServerThe easy sibling of the Shared-Skills server: the same team snippet library as a remote MCP server, but auth is one shared token you paste into a header — no OAuth, no consent page, no signing-key migration. One Supabase Edge Function, one RLS-locked table, live in ~15 minutes.
  15. advanced · 7 stepsMulti-Agent A2A System on Cloud RunThree ADK agents — orchestrator, researcher, writer — each deployed as its own Cloud Run service and wired together with Google's open Agent2Agent (A2A) protocol.
  16. beginner · 9 stepsBuild a Context-Aware RAG App on Google CloudDocument Q&A end-to-end on GCP — Cloud SQL Postgres with pgvector, Vertex AI for embeddings and Gemini, FastAPI on Cloud Run. Task-typed embeddings, metadata-filtered retrieval, and optional contextual chunking. IAM auth all the way down, no passwords.
  17. beginner · 6 stepsDeploy Gemma as an Open API on Cloud RunServe Google's open Gemma 4 on a serverless Cloud Run GPU — one Dockerfile, one deploy command, unauthenticated behind a single HTTPS URL. Then call it straight from a static webpage with streaming, no backend of your own.
  18. intermediate · 7 stepsGive Your Agent Long-Term MemoryBuild the memory layer that survives the session-15 test — a local SQLite fact store with temporal validity, LLM fact extraction, rolling summaries, and embedding recall, all offline via Ollama. Companion build to the Agent Long-Term Memory drip.
  19. advanced · 8 stepsTrain a Reasoner with GRPOTeach a small model to reason with RLVR — GRPO against a verifiable math reward using Unsloth + TRL, LoRA on a single GPU, with pass@1 measured before and after so you can watch the accuracy line move. Companion build to the RLVR & Process Rewards drip.
  20. intermediate · 7 stepsQuantize & Run a ModelTake one Hugging Face model, quantize it two ways — GGUF (llama.cpp) and AWQ — benchmark VRAM, speed, and quality side by side, then serve the winner via Ollama or vLLM. Companion build to the Quantization drip.
  21. beginner · 7 stepsIn the queueCreate Your Business Website with AIGo from idea to live website using AI — no coding required. Design, copy, and deploy in one sitting.
  22. beginner · 8 stepsIn the queueAutomate Your Workflow with AIConnect your tools and let AI handle the busywork — email triage, data entry, and report generation using Make or Zapier.
  23. beginner · 6 stepsIn the queueBuild Your Personal AI AssistantCreate a custom AI tailored to your job — with your documents, your tone, and your processes baked in.
  24. beginner · 7 stepsIn the queueLaunch a Customer Support ChatbotDeploy an AI chatbot that answers customer questions using your FAQ and docs — live in under an hour.
  25. beginner · 6 stepsIn the queueTurn Your Docs into a Searchable AI BrainUpload your company docs or notes and create an AI-powered search interface anyone on your team can use.
  26. beginner · 7 stepsIn the queueAI-Powered Content PipelineSet up a system that drafts blog posts, social media, and newsletters from your ideas — with your voice and brand.
  27. beginner · 9 stepsIn the queuePrompt Engineering That Actually WorksMaster the frameworks that get consistently great output — structured prompts, chain-of-thought, and few-shot patterns applied to real tasks.
  28. intermediate · 8 stepsIn the queueCreate an AI-Powered Slack BotAdd an AI teammate to your Slack workspace that answers questions, summarizes threads, and triggers workflows.
  29. beginner · 7 stepsBuild Your AI-Powered NewsletterCurate content from your favorite sources and let AI draft, format, and schedule a newsletter — no writing block, no coding.
  30. beginner · 6 stepsCreate an AI Study BuddyUpload your notes and textbooks and get AI-generated flashcards, practice quizzes, and personalized study plans.
  31. beginner · 8 stepsLaunch an AI Social Media ManagerGenerate on-brand posts, suggest hashtags, plan a content calendar, and schedule across platforms — all with AI.
  32. beginner · 6 stepsBuild Your AI Meeting AssistantTranscribe meetings, extract action items, generate summaries, and send follow-up emails — automatically.
  33. beginner · 7 stepsDesign Your AI-Powered PortfolioLet AI write your copy, suggest layouts, optimize for SEO, and turn your projects into a polished personal site.