Kilo CLI — The Terminal-Based Agentic Coding Tool You Need in 2026
What is Kilo CLI? If you spend a significant portion of your day living in the terminal, Kilo CLI might be the most important tool you add to your workflow this year. It is an open-source, terminal-based agentic coding platform built on OpenCode (MIT-licensed), designed...
FastAPI + LLM
The previous post covered why FastAPI fits GenAI applications well in principle. This one is the practical follow-through: a closer, more complete look at actually wiring an LLM into a FastAPI application — from a single endpoint to a more realistic setup with conversation...
AgentScope Model Integration
The best model for a given task today may not be the best choice in six months — pricing changes, new models launch, and capabilities shift. A framework that locks an agent's logic to one specific provider makes that evolution expensive to keep up with. AgentScope was built...
Agent Memory Architecture
An agent working through a long, multi-step task needs to remember what it's already tried, what it's learned, and what still needs doing — but an LLM's context window, as covered in the tokens post earlier in this series, is finite. Agent memory architecture is the set of...
Building an AI CRM Agent
A sales or support team's CRM holds a huge amount of valuable, structured information — contacts, deal stages, communication history, ticket status — but keeping it accurate and acting on it consistently is a constant manual burden. An AI CRM agent applies the agent pattern...
Building a Coding Agent
Coding is one of the domains where AI agents genuinely shine — not because writing code is easy for an LLM, but because code has something rare among agent tasks: a built-in, objective way to check if the work actually succeeded. Tests pass or they don't; code runs or it throws …
Autonomous AI Agents
Some AI agents wait for a human to approve every action. Others run for hours, making dozens of decisions on their own, only checking in when something genuinely needs human judgment. That second category — agents operating with minimal ongoing human oversight — is what's...
Agent Tools
An LLM without tools can only describe what it would do. Give it tools, and it can actually do it — search the web, query a database, run code, send a message. Tools are what turn an agent's reasoning into real-world action, and designing them well is one of the most …
AI Agent vs LLM
It's easy to use "AI agent" and "LLM" interchangeably in casual conversation, but they describe genuinely different layers of a system. An LLM is a component — a powerful one — while an agent is an entire architecture built around that component, adding the pieces that let it...
What is an AI Agent?
A chatbot answers your question and stops. An AI agent, by contrast, can decide what to do, take an action, observe the result, and decide what to do next — repeating that cycle until it's actually accomplished a goal. This shift, from generating a single response to...
AI Agent Directory - Few Shots LLM Models
The AI agent ecosystem is growing fast. Here's a quick directory of notable AI startups and a couple of few-shot LLM models worth knowing about. Two lines each — just enough to know what they do and why they matter. AI Agent Directory (Test) Can of Soup — An AI-powered app that …
Reducing RAG Hallucinations
Introduction: When RAG Lies Despite the Evidence Retrieval-Augmented Generation was supposed to solve hallucinations. By grounding the model in retrieved documents, we assumed it would stick to the facts. But RAG systems hallucinate too. They ignore retrieved context and...
RAG vs Fine-Tuning
Introduction: Two Paths to Customization When organizations want an LLM to perform well on their specific domain, they face a strategic choice between two fundamentally different approaches. Retrieval-Augmented Generation augments the model's context at inference time by...
What is RAG?
Introduction: The Knowledge Gap in Large Language Models Large Language Models are trained on vast corpora of internet text, books, and code. They can write poetry, debug software, and explain quantum mechanics. But they have a critical limitation: their knowledge is frozen...
LangChain Runnable Architecture
The Evolution to LCEL LangChain has evolved significantly since its early days. The original API relied heavily on explicit chain classes like LLMChain and SequentialChain. While functional, these classes were sometimes rigid and required developers to learn specific APIs for...
LangChain Output Parsers
The Structured Output Problem Large Language Models generate text. Production systems consume structured data. This fundamental mismatch is one of the most persistent challenges in building reliable GenAI applications. When you ask a model to return a JSON object, it might...
LangChain Messages
The Message Paradigm Modern conversational AI is built on a message-based interaction model. Unlike early text completion systems that processed raw strings, today's chat models are trained on structured conversations where each utterance has a specific role. LangChain...
LangChain Prompts
Beyond String Concatenation Prompting is the primary interface for controlling LLM behavior. In simple scripts, it is tempting to construct prompts using Python f-strings or basic string formatting. However, this approach quickly becomes unmanageable in production...
LangChain Models
The Model Abstraction At the heart of every LangChain application is a language model. However, LangChain does not implement its own models. Instead, it provides a unified interface that wraps models from dozens of providers. This abstraction is one of the framework's most...
LangChain Architecture Explained
Layered Design Philosophy LangChain's architecture is deliberately layered, resembling the design of modern web frameworks. Each layer has a specific responsibility, and developers can interact with the framework at whatever level of abstraction suits their needs....