AI Summary Hub

Changelog / Updates

Recent updates and new topics in AI Summary Hub.

2026-05 — Quality pass and frameworks index

  • Added ## Sources sections (3–4 seminal papers/official docs per article) to all en/agents/*.mdx and en/reasoning-patterns/*.mdx files.
  • New en/frameworks/index.mdx — PyTorch vs TensorFlow overview with when-to-use table, mermaid diagram, and sources.
  • Added When to use / When NOT to use table to en/prompt-engineering/index.mdx.
  • Updated intro.mdx article count to reflect current scale (150+ articles, 50+ categories).
  • Glossary expanded with 2025–2026 terms: MCP, A2A, DPO, GRPO, RLHF, Agentic RAG, GraphRAG, Context Engineering, Computer Use, Extended compute, Vibe coding, LLM-as-judge.
  • API key hygiene: en/model-providers/openai.mdx, deepseek.mdx, and mistral.mdx updated to read credentials from environment variables.

2026-01 — Agents expansion (Batch 3)

  • New agent articles: a2a-protocol.mdx, agent-runtimes.mdx, dag-agents.mdx, planner-executor.mdx, self-critique.mdx, security.mdx, debugging.mdx, evaluation.mdx, conversational-memory.mdx, memory.mdx, tools-actions.mdx, prompt-engineering.mdx.
  • New en/agents/ sub-pages covering A2A Protocol (Linux Foundation, 150+ org adoption), AgentsKit, Hermes Agent, and Paperclip runtimes.
  • Full i18n retranslation batch 3 — 244 files across de, es, fr, pt-BR, zh-Hans locales.

2025-Q4 — MLOps, Prompt Engineering, Claude Code, MCP, RAG

  • MLOps category: CI/CD (dvc.mdx, model-registry.mdx), data engineering (airflow.mdx, kafka.mdx, spark.mdx), deployment (kubeflow.mdx, ml-kubernetes.mdx), and monitoring sub-sections.
  • Prompt Engineering expansion: prompt-formats.mdx, automatic-prompt-engineering.mdx, debiasing-techniques.mdx, prompt-ensembling.mdx, self-evaluation-calibration.mdx, step-back-prompting.mdx, structured-outputs.mdx, system-role-contextual-prompting.mdx, temperature-top-k-top-p.mdx, max-tokens-stop-sequences.mdx, self-consistency.mdx.
  • Claude Code category: index.mdx, claude-md.mdx, context-management.mdx, mcp-plugins.mdx, prompt-caching.mdx, skills.mdx, thinking-modes.mdx.
  • MCP (Model Context Protocol): index.mdx, building-clients.mdx, building-servers.mdx.
  • RAG expansion: agentic-rag.mdx, graphrag.mdx added alongside existing architecture, embeddings, vector-databases, and examples articles.
  • LLMs post-training: post-training.mdx covering RLHF, DPO, GRPO.
  • Model providers deep-dives: Anthropic, OpenAI, Google Gemini, Meta Llama, Mistral, DeepSeek, Cohere.
  • Vibe coding article added.

2025-Q3 — Core content expansion (Batch 1 & 2)

  • Agents category: index.mdx, multi-agent-systems.mdx, frameworks-overview.mdx, crewai.mdx, autogen.mdx, langgraph.mdx, anthropic-tool-use.mdx.
  • Reasoning patterns: cot.mdx, react.mdx, tot.mdx, rdd.mdx.
  • Frameworks: pytorch.mdx, tensorflow.mdx.
  • Case studies added: ChatGPT, DALL·E, Claude, Gemini, BART, Grok, DeepSeek, Qwen.
  • Edge AI articles: ONNX, PyTorch Mobile, TFLite.
  • i18n initial pass for es, pt-BR, de, fr, zh-Hans locales.

Initial release

  • Wiki structure with core topic placeholders.
  • Core docs for RAG, Transformers, LLMs, Agents, Subagents, RDD.
  • 6 locales (en, es, pt-BR, de, fr, zh-Hans).
  • Dark default theme, Algolia search, GitHub Pages deployment.

Watch the repo for new topics and translations.

On this page

2026-05 — Quality pass and frameworks index2026-01 — Agents expansion (Batch 3)2025-Q4 — MLOps, Prompt Engineering, Claude Code, MCP, RAG2025-Q3 — Core content expansion (Batch 1 & 2)Initial release