The tools you actually need, and the ones you don’t.
A curated guide to the platforms, frameworks, and infrastructure that make AI deployments work in practice.
Tool Guides
- n8n — Workflow automation and AI agent orchestration
- Make — Visual automation; cloud-only, pay-per-action vs n8n’s pay-per-execution
- Ollama — Local open-source LLM serving
- vLLM — Production-scale inference
- Zapier — No-code automation; cloud-only, per-task pricing
- Open WebUI — Browser chat interface for self-hosted LLMs
- LiteLLM — Unified API gateway with fallback and cost routing
- Langfuse — LLM observability, cost tracking, and evaluation
- Flowise — Low-code agent builder for RAG pipelines and multi-agent workflows
- Langflow — Visual builder for LangChain agents; pairs with Flowise
- LangChain — Framework for LLM apps; LangSmith observability, LangGraph agents
- Mistral AI — Open-weight models for self-hosted or API use
- Azure AI Foundry — Enterprise AI platform; compliance-ready, lock-in tax
Core Concepts
- Self-Hosted AI — When to run models on your own hardware
- Vector Databases — The storage layer for retrieval
- MCP — The protocol for tool connections
Layer Hubs
- Integration Layer — MCP, APIs, workflow tools
- Infrastructure Layer — Self-hosting vs cloud
Each tool guide will include: what it does, what it costs, where it breaks, and how to recover.
Latest Updates
- June 2026 — LangChain
- June 2026 — Azure AI Foundry
- June 2026 — Mistral AI
- June 2026 — Langflow
[All articles in Stack & Tools →](All articles in Stack & Tools)