AI Migration & Build Tutorials
Every tutorial here is written by someone who actually shipped the migration or pattern in production — with the exact code, the failure modes nobody warns you about, and a rollback path if it goes sideways at 2am.
More tutorials are landing every week. If a migration you're planning isn't here yet, the most popular destinations live on the head-to-head and cost calculator hubs — start there.
18 pages · updated 2026
Agent Eval with Langfuse (2026): Traces, Datasets, Scoring, Replay
5-step eval pipeline for LLM agents with Langfuse: trace collection, dataset creation, LLM-as-judge scoring, dashboards, and trace replay for regression testing. Sourced from langfuse.com/docs.
ReadAudit Trail for LLM Prompts (SOC 2, 2026): Build Guide
How to build a SOC 2-grade audit trail for LLM prompt invocations in 2026. Log schema, storage choices (CloudWatch / Cloud Logging / Datadog), retention, encryption, access controls, query patterns, and the AI-specific evidence auditors expect. Code patterns in Python and TypeScript.
ReadBuild a 3-Agent CrewAI Crew (2026): Researcher, Writer, Editor
Step-by-step tutorial to build a 3-agent CrewAI crew in 2026 — researcher, writer, editor agents with tools, tasks, and a manager. Code-first, production-ready. Sourced from docs.crewai.com.
ReadBuild a ReAct Agent with LangGraph (2026): 5-Step Tutorial
Step-by-step tutorial to build a production-ready ReAct agent with LangGraph in 2026. State graph, tool nodes, conditional edges, memory, and streaming. Code-first. Sourced from official docs.
ReadBuild GraphRAG (2026): Microsoft GraphRAG v2 + Knowledge Graph + Claude Generation
When and how to build GraphRAG: Microsoft GraphRAG v2 pipeline, entity extraction, Leiden community detection, global vs local search, costs ($100-500/1M tokens corpus), LightRAG comparison. Full Python code.
ReadBuild an LLM Red-Team Suite (2026): Promptfoo + Garak + Inspect AI
End-to-end tutorial: build a production-grade LLM red-team suite in 2026 using Promptfoo (eval + red-team), NVIDIA Garak (vulnerability scanner), and UK AISI's Inspect AI (rigorous evaluation framework). CI integration, observability, audit-grade reporting.
ReadBuild RAG with pgvector (2026): Postgres + HNSW + Claude Sonnet 4.6
End-to-end tutorial: pgvector 0.7+ on Postgres 16, HNSW index, OpenAI embeddings, hybrid BM25+vector search, Supabase/Neon managed options, Claude generation. No separate vector DB required.
ReadBuild RAG with Pinecone (2026): Serverless Index + Claude Sonnet 4.6
End-to-end Python tutorial: Pinecone v6 serverless index, text-embedding-3-small, Claude Sonnet 4.6 generation, metadata filtering, namespaces, and streaming. Production-ready, no magic wrappers.
ReadBuild a SaaS With Bolt.new (2026): 6-Phase Tutorial With Worked Prompts
End-to-end tutorial for building a real SaaS with Bolt.new in 2026. Pick the right plan, write a spec-shaped prompt, iterate on the generated app, add Supabase + Stripe, deploy via Netlify, then graduate to your own repo when you outgrow Bolt. 3 worked prompts. Sourced from bolt.new docs + stackblitz.com.
ReadClaude Code Best Practices (2026): The Canonical Playbook With Hooks, Skills, MCP
The definitive 2026 playbook for Claude Code (Anthropic's CLI agent). Install + auth, CLAUDE.md + AGENTS.md memory, skills vs slash commands, hooks (PreToolUse/PostToolUse/Stop), subagents + parallel tool use, MCP servers (filesystem/github/slack/playwright/web-fetch), cost discipline by model tier, common antipatterns. Real config examples. Sourced from docs.claude.com.
ReadConfigure OpenAI Zero Data Retention (2026): Step-by-Step
How to configure OpenAI's Zero Data Retention (ZDR) for an Enterprise project in 2026 — eligibility, dashboard navigation, project key isolation, eligible-endpoint verification, audit-trail checks. Plus the workflow for combining ZDR with BAA for HIPAA workloads.
ReadCursor Rules for Next.js (2026): .cursor/rules/ Tutorial With Real .mdc Examples
Definitive 2026 guide to Cursor Project Rules for Next.js 16 — .cursor/rules/*.mdc files with YAML frontmatter, globs, alwaysApply. Full .mdc examples for App Router (Server Components, server actions, Cache Components), Tailwind/shadcn, database/auth, testing, and sharing rules across teams. Sourced from cursor.com/docs/context/rules.
ReadHybrid Search: BM25 + Dense Retrieval (2026) — RRF, Qdrant, Weaviate, Cohere Rerank
Why hybrid beats either-alone by 5-12 NDCG@10 on BEIR. Reciprocal Rank Fusion in 10 lines of Python. Qdrant/Weaviate native hybrid + Cohere rerank-v3.5 pipeline. Full production code.
ReadImplement DLP for LLM Apps (2026): PHI / PII Redaction Before Inference
How to implement Data Loss Prevention for LLM applications in 2026 — pre-LLM PHI/PII detection and redaction, tooling choices (Microsoft Presidio, AWS Comprehend Medical, Nightfall, Skyflow, Google Cloud DLP), de-identification per HHS Safe Harbor, audit logging, and production patterns.
ReadMigrate from GitHub Copilot to Cursor (2026): 5-Phase Tutorial With Settings Import
Step-by-step migration from Copilot to Cursor in 2026. Audit your Copilot usage to right-size the Cursor plan, install Cursor, import VS Code settings/extensions/keybindings, configure Composer + Agent + MCP, then cancel Copilot OR run the $30/mo stack. With code, conflict-list, and rollback plan. Sourced from cursor.com/docs.
ReadMulti-Agent Coordination Patterns (2026): Supervisor, Swarm, Hierarchical, Network
Deep dive into multi-agent coordination patterns in 2026: supervisor, swarm, hierarchical, and agent network. LangGraph and AutoGen implementations, cost math, and when to use each. Sourced.
ReadRun Claude with EU Data Residency (2026): The Three Paths
How to run Anthropic Claude with EU data residency in 2026 — three paths: AWS Bedrock in eu-west-1 / eu-central-1, Google Vertex AI in europe-west1+, and Anthropic Enterprise direct EU residency. Configuration, contracting, and the GDPR + EU AI Act implications.
ReadOpenAI to Claude Migration Guide (2026): SDK Diffs + Cost Delta Calc
Step-by-step migration from OpenAI to Anthropic Claude in 2026. SDK code diffs (chat completion → messages.create), system-prompt restructuring, prompt-caching setup, tool-use schema changes, and a worked cost-delta calculator showing real $ savings. Production-ready, no marketing fluff.
Read
Stop guessing your AI bill.
Digital Dashboard Hub turns your real spend across OpenAI, Anthropic, and Google into one live dashboard — usage, cost, budget alerts, model mix. 14 days free.
Try DDH free