A chain runs start to finish or it doesn't run at all. LangGraph exists because most real agent workflows need a third option: pause, wait, and pick back up exactly where they left off. ........ LangGraph models an agent as a directed graph: nodes are functions that read and update a typed state object, edges define execution order and branching. A checkpointer saves a snapshot of that state after every node executes, organized by thread — the mechanism behind resuming a fail

LiteLLM does the same job OpenRouter does — one API for every LLM provider — but hands you the router itself instead of renting you access to one. ........ LiteLLM ships as two things: a Python SDK for routing and cost tracking inside your own code, and a Proxy Server that runs as a standalone gateway with an OpenAI-compatible endpoint. Five routing strategies are available: simple-shuffle (the default), least-busy, latency-based, usage-based, and cost-based — the last two ne

Teable is an open-source, PostgreSQL-backed spreadsheet-database hybrid — an Airtable-style grid interface sitting on a real SQL backend, with AI fields, natural-language automations, and an app builder layered on top. Version 3.0's headline addition is the Teable Skill: external AI agents, including Claude Code and ChatGPT, can now connect directly to a Teable Base and manage its tables, fields, views, records, automations, and apps, rather than going through a generic Web A