Category
Step-by-step AI tooling guides with the gotchas other tutorials skip.
56 articles
ai-tutorialsBuild a Python scraper that drives Playwright, screenshots a rendered page, sends the image to Claude Sonnet 4.6, and validates the output with Pydantic.
ai-tutorialsEnd-to-end: wire Claude into GitHub Actions to comment AI-generated review notes on every pull request, with secrets, cost estimates, and security guardrails.
ai-tutorialsUpload a CSV, ask Claude to analyse it with pandas in the sandbox, retrieve the generated chart, and validate cost-per-query — full Python walkthrough.
ai-tutorialsPaste a claim from your paper, the agent queries the arXiv API, ranks supporting and contradicting work with Claude Sonnet 4.6, and prints a BibTeX file.
ai-tutorialsHands-on walk-through of an agent loop with Claude tool-use, file/bash tools, a sandbox, and a failing-test exercise comparing against Cursor and Claude Code.
ai-tutorialsRegister a Discord app, wire discord.py to Claude tool-use with web-search and calculator tools, deploy on Railway's free trial — full Python walkthrough.
ai-tutorialsWire LangChain, langchain-anthropic, and ChromaDB into a Streamlit chat UI that answers questions over uploaded PDFs, with retrieval evaluation included.
ai-tutorialsBuild a multimodal RAG app end-to-end in Python: LangChain + ChromaDB with OpenCLIP, query a product catalogue with photos or text, answer with Claude Sonnet.
ai-tutorialsWire Claude tool-use to the Google Calendar API in Python with list_events, create_event, and find_free_slot tools — full OAuth-to-agent walkthrough.
ai-tutorialsEnd-to-end tutorial: scaffold a Bolt for Python Slack app, route every channel message through Claude for safety classification, ship to Fly.io for production.
ai-tutorialsEnd-to-end Python tutorial: Pinecone serverless index, 10-doc PDF knowledge base, Claude Sonnet tool-use, FastAPI endpoint, evaluated on 5 support queries.
ai-tutorialsFollow-along tutorial: scaffold a Gradio Blocks chat UI that calls Claude via the Anthropic Python SDK, push to a free-tier Space, wire env-var secrets, test the URL.
ai-tutorialsEnd-to-end tutorial: scaffold a FastAPI Anthropic-API app, add a KV-backed rate-limit cache, deploy via Wrangler. Honest about Workers-Python beta status.
ai-tutorialsTrain an SDXL LoRA on a custom subject in 60-90 minutes on a 16GB GPU. Dataset, kohya_ss / diffusers training, inference, 5-prompt evaluation.
ai-tutorialsEnd-to-end Python tutorial: load DistilBERT SST-2 with Transformers, expose a /analyze endpoint via FastAPI lifespan, batch inputs, Dockerize, ship to Fly.io.
ai-tutorialsPipe a Wikipedia paragraph through Claude tool-use to extract subject-predicate-object triples, build a NetworkX DiGraph, and visualise it with PyVis.
ai-tutorialsWalk-through: install DeepEval, build a 20-sample dataset, wire G-Eval + hallucination metrics, trace via LangSmith, and gate a model upgrade on the eval.
ai-tutorialsBootstrap Next.js 15 App Router, add Supabase email magic-link auth, gate /chat behind a session, stream Claude responses, and deploy to Vercel free tier.
ai-tutorialsEnd-to-end walkthrough: write a Python MCP server that exposes an Obsidian vault read-only to Claude Desktop, then ask Claude to summarise, connect, and draft notes.
ai-tutorialsBuild a Python pipeline that polls RSS feeds, filters items by Claude against your interests, renders a Markdown digest, emails via Resend, and runs on cron.
ai-tutorialsBuild a Python agent that scrapes Amazon prices with Playwright, classifies deals with Claude, emails alerts via Resend, and runs on a Fly.io cron.
ai-tutorialsSpin up Upstash or local Redis, wire LangChain RedisCache and RedisSemanticCache, measure latency plus token-cost reductions on 50 sample queries, then tune TTL.
ai-tutorialsEnd-to-end QLoRA fine-tune of Mistral-7B-Instruct on a 1,000-row instruction dataset, on a single 16GB GPU, with the adapter pushed to Hugging Face Hub.
ai-tutorialsWire Semantic Scholar's Recommendations API to Claude Sonnet 4.6 in a Streamlit app: paste a paper, get a ranked list of similar work with three-sentence summaries.
ai-tutorialsSpin up Postgres with pgvector, generate Cohere or OpenAI embeddings, build an HNSW index, run k-NN queries, and layer metadata filters — production-shaped.
ai-tutorialsForce Claude Sonnet to return validated JSON via tool_use, parse with Pydantic v2, stream with retries — extract structured orders from email prose.
ai-tutorialsWire Claude tool-use to a Chinook Postgres database, generate safe SELECT-only queries from natural language, and render Markdown result tables — Python walkthrough.
ai-tutorialsWire WhatsApp Cloud API to Claude through a FastAPI webhook on Render's free tier — token, webhook verification, message echo, deploy, send-and-reply test.
ai-tutorialsPython CLI that fetches a YouTube transcript via youtube-transcript-api, summarises with Claude Sonnet, outputs Markdown with timestamps — packaged for PyPI.
ai-tutorialsWire a FastAPI webhook receiver to Claude for urgent/billing/general/spam classification, verify Intercom HMAC signatures, and route conversations via the Admin API.
ai-tutorialsWire Claude into Git's prepare-commit-msg hook so every git commit opens with a Conventional Commits draft generated from your staged diff, plus a clean opt-out flag.
ai-tutorialsExtract text from a PDF, prompt Claude to draft 10-30 Anki-style flashcards with tags, then export a ready-to-import .apkg deck via genanki.
ai-tutorialsWire boto3, Claude vision, and psycopg into an end-to-end pipeline that tags every S3 upload and serves a /search?tag= API in FastAPI.
ai-tutorialsBrowser-only voice chatbot — mic input via Web Speech API, streamed Claude responses, Web Audio playback. Pure HTML/JS deployed on GitHub Pages.
ai-tutorialsEnd-to-end Python project: a receipt photo, Claude vision extracts items, prices, tax, totals; Pydantic validates; SQLite stores; CSV exports.
ai-tutorialsEnd-to-end Python tutorial: a Slack app that posts a daily standup prompt, collects replies, then asks Claude to digest blockers and status by member.
ai-tutorialsWire Weaviate Cloud hybrid BM25 search, Cohere embed-v4.0 vectors, and rerank-v3.5 into a FastAPI /search endpoint over 100 Wikipedia chunks.
ai-tutorialsEnd-to-end Smolagents tutorial. Wire Hugging Face's CodeAgent to a local Qwen 3 on Ollama, add two custom arXiv tools, and ship a working research agent on a laptop.
ai-tutorialsInstall vLLM 0.20.1, serve Llama 3.1 8B, scale across GPUs with tensor parallelism, and pick FP8 vs W8A16 by hardware. A 90-minute walkthrough for ML engineers.
ai-tutorialsHands-on autogen-agentchat walkthrough (v0.7.x, May 2026): install, wire two AssistantAgents into a research crew, inspect the conversation. Status: maintenance mode.
ai-tutorialsInstall Promptfoo, write a YAML config, run side-by-side evals on GPT-5 vs Claude Sonnet 4.5, and ship the report. A 60-minute walkthrough for devs.
ai-tutorialsAn AI agent acts. A chatbot replies. An assistant suggests. Why the distinction matters for dev teams choosing what to build and how to test it.
ai-tutorialsWire Claude Desktop to filesystem, GitHub, and fetch MCP servers in 30 minutes. Config-file walkthrough, token scopes, and the pitfalls devs hit first.
ai-tutorialsA 60-minute Python walkthrough that builds a Researcher + Writer + Editor crew with CrewAI, picks the right Process mode, and ships structured output.
ai-tutorialsStep-by-step ElevenLabs Python tutorial: signup, voice picker, Turbo and Eleven v3, WebSocket streaming, Hindi and Tamil. Around 50 lines of code.
ai-tutorialsThree different approaches solving three different problems. A plain-English decision tree for picking the right one — and avoiding the expensive default.
ai-tutorialsWrap Claude, GPT-5.5, or a local Ollama model in a streaming chat UI using Gradio 5.x and ~30 lines of Python. Tutorial includes Hugging Face Spaces deploy.
ai-tutorialsA 60-minute Qdrant tutorial for dev teams: Docker install, sentence-transformers embeddings, hybrid search with metadata filters, and a FastAPI endpoint.
ai-tutorialsRAG in 60 minutes on a laptop: LangChain + ChromaDB, in-process vector store, free local embeddings, Claude or GPT for answers. No Docker, no Pinecone.
ai-tutorialsA step-by-step Smolagents walkthrough in Python. Roughly 40 lines of code, any LLM, working code-writing agent in 30 minutes — no LangChain, no LangGraph.
ai-tutorialsBuild a Streamlit LLM-pricing dashboard: load OpenAI + Anthropic spend logs, chart cost-per-day with Altair, add filters, deploy free on Community Cloud.
ai-tutorialsAttention is the mechanism in a transformer where each token decides how much to look at every other token. Three matrix products, one softmax, weighted sum.
ai-tutorialsMixture-of-Experts splits a transformer's feed-forward layers into experts and activates only a few per token. The result: cheaper inference for dev teams.
ai-tutorialsRAG explained for devs in 2026: when retrieval-augmented generation is the right pattern, when long-context beats it, and the failure modes teams hit first.
ai-tutorialsRLHF turned GPT-3 into ChatGPT by aligning models to human preferences. By 2026, simpler methods like DPO have largely replaced it for new alignment work.
ai-tutorialsStep-by-step tutorial: build a working Model Context Protocol server in TypeScript, expose it to Claude Desktop, and learn the gotchas other guides skip.
No articles match the selected topic yet. Clear filter .