Agents
An AI agent is an autonomous system powered by a Large Language Model (LLM) that can perceive its environment, make decisions, plan, and take actions to achieve a specific goal.
Agents Sections
- Toolkits (10)
- Key Concepts: Codex App Server Bidirectional JSON-RPC Conversation Primitives (Item, Turn, Thread) Agent Client Protocol (ACP)
OpenAI's Codex App Server decouples agent logic from UI using a bidirectional JSON-RPC protocol. Learn about its core primitives and why it diverges from MCP.
- Key Concepts: Vector Embeddings Semantic Search Hybrid Search HNSW Indexing pgvector
An overview of vector databases, which are specialized systems designed for efficient similarity searches in high-dimensional spaces. Covers core algorithms like HNSW and IVF, the recall-latency trade-off, and their critical role in applications like RAG and semantic search.
- Key Concepts: AI social media agent writing style analysis persistent memory Nebius AI Composio ScrapeGraph Memori Streamlit
A technical guide to building an AI social media agent that scrapes your viral tweets, learns your writing style with persistent memory, and posts autonomously via Composio and Nebius AI.
- Key Concepts: Claude Agent SDK Agentic Feedback Loop Subagents Model Context Protocol (MCP) Verification Loops Tool Design Context Management
A practical guide to developing AI agents with the Claude Agent SDK, covering core principles like the agentic feedback loop, subagents, tool design, and verification.
- Key Concepts: Deep Agents Context Compression Context Rot Filesystem Offloading Summarization Needle-in-the-Haystack Test Goal Drift
A technical guide to the context compression techniques (offloading, summarization) used in the LangChain Deep Agents SDK to manage long-running tasks.
The performance of an AI agent is directly linked to the quality of its tools. This guide details the essential principles for designing robust and reliable functions for agents, covering best practices like clear naming, descriptive docstrings, specific functionality, clean outputs, and informative error handling to maximize effectiveness.
- Key Concepts: Agentic Workflow Reasoning Loop ReAct Framework Orchestration Agent Tools Observability
Modern AI agents do more than simply respond to prompts; they execute workflows. A workflow is a multi-step sequence of actions that an agent performs autonomously to achieve a complex goal.
- Key Concepts: Context Engineering Context Window Lost-in-the-Middle Multi-Agent Patterns Memory Systems LLM-as-a-Judge
A reference guide on Context Engineering for AI agents, detailing the principles of managing a language model's context window to maximize effectiveness. It covers foundational skills, architectural patterns, and operational best practices.