As agents mature, more and more products will need to serve two kinds of users at once: humans and agents. A CLI is one of the easier forms for an agent to call, and in the future many products may want to ship their own CLI to expose their core capabilities to agents. This post walks through the design of a CLI we recently open-sourced for a platform product, covering six areas: language choice, login and authorization, command and flag design, skills, distribution, and security.

Read more

Self-evolution is a core module of the Agent Harness. With it, an Agent can keep improving across long-running tasks: refining its own skills, recording user feedback and preferences, and reviewing its own work to keep getting better. This post uses the open-source project CowAgent as a worked example to walk through the architecture and engineering behind a five-layer self-evolution mechanism.

Read more

Most public reviews of DeepSeek V4 focus on chat or single-turn coding. This post documents how deepseek-v4-flash behaves inside CowAgent's full agent loop — planning, complex coding, long-term memory, browser control, knowledge-base construction, and very long documents — across six end-to-end tasks. The goal is to give readers a sense of what to expect when running V4 Flash as an agent's default model, including where it holds up and where it occasionally drops a fine-grained constraint.

Read more

RAG (Retrieval-Augmented Generation) has developed rapidly over the past two years. It has moved past the early proof-of-concept stage and into a critical period of large-scale enterprise adoption. Along the way, simple document chunking and basic vector search are no longer enough to meet enterprise expectations for knowledge base Q&A quality, and every stage of the RAG pipeline needs deeper optimization.

Read more

AgentMesh is an open-source multi-agent framework. Its core goal is to solve communication and collaboration between multiple agents, so that together they actually achieve "1+1>2". It helps users quickly build their own multi-agent teams, or gives several existing standalone agents the ability to work together, so they can take on more complex tasks.

Read more

The DeepSeek-R1 LLM has strong reasoning capabilities and shows major gains on math, code, and natural language reasoning tasks. Official and third-party online services are all unstable to some degree, and data security and privacy are a concern, so self-hosting the open-source DeepSeek model is a good option for individuals and companies alike. This post walks through deploying and testing the full-parameter deepseek-r1-671b model. The same process also applies to deepseek-v3-671b and to the smaller versions.

Read more

RAG (Retrieval-Augmented Generation) is one of the mainstream techniques behind today's AI agent applications. It addresses the shortcomings of LLMs in Q&A scenarios (limited knowledge, outdated knowledge, hallucination, and so on), and makes it possible to deploy LLMs in specialized domains, especially in enterprise settings, where they have to meet real production needs and business requirements. This post uses the evolution of the LinkAI knowledge base as an example to walk through RAG optimization in practice.

Read more

ChatGPT has recently taken the internet by storm with its strong conversational and information-synthesis abilities. It can write code, revise papers, and tell stories; there is almost nothing it can't do. That led me to a bold idea: could its conversational model turn our WeChat into a smart bot that gives unexpected replies when chatting with friends, so we never again have to worry about our girlfriends keeping us from gaming work?

This project is a ChatGPT-based WeChat chatbot. It generates replies through the OpenAI API, uses itchat to receive WeChat messages and reply automatically, and supports LinkAI for agent capabilities.

Read more

An operating system is, at its core, just a program. Like the applications we write, it ends up as binary machine instructions executed by the CPU; as computers evolved, this particular program simply took on more special responsibilities. Today we set aside the many functions of an OS and go back to what it is as a program: the first step is to get its first line of code running on a computer.

Read more

Preface

The operating system is the foundational software behind almost everything developers do. Every application we write runs on top of one, and as soon as you need to optimize an application's performance or resource usage, you need to understand how the OS works. The OS is also the entry point to learning low-level computing, and understanding its internals quietly improves the way you write programs.

Read more