Reading
machine-core is a flexible framework for building AI Agents, supporting MCP (Model Context Protocol) integration, dynamic OpenAPI tool generation, vector-based RAG tool filtering, and multiple LLM and Embedding providers.
Section 01
machine-core is a flexible framework for building AI Agents, supporting MCP (Model Context Protocol) integration, dynamic OpenAPI tool generation, vector-based RAG tool filtering, and multiple LLM and Embedding providers.
Section 02
machine-core is a flexible framework specifically designed for building AI Agents. Its core positioning is to provide a clear, scalable infrastructure that allows developers to quickly build intelligent agent systems with tool calling, knowledge retrieval, and multi-model support capabilities.
Unlike traditional Agent frameworks, machine-core emphasizes clear separation of architecture: the infrastructure layer (AgentCore) is decoupled from the execution mode layer (BaseAgent), enabling developers to choose the appropriate Agent type based on specific scenarios without being restricted by the framework's preset structure.
Section 03
machine-core has built-in support for 7 LLM providers and 3 Embedding providers:
LLM Providers:
qwen3-vl:32b)gpt-4o-2)grok-2-latest)llama-3.3-70b-versatile)Embedding Providers:
nomic-embed-text)text-embedding-3-large)This multi-provider architecture allows developers to easily switch models via environment variables or configurations, enabling flexible model routing and fallback strategies.
Section 04
MCP is an open protocol proposed by Anthropic, aiming to standardize the interaction between AI models and external tools/data sources. machine-core natively supports MCP:
This means you can quickly integrate any MCP-compliant tools (such as file system access, database queries, API calls, etc.) into your Agent.
Section 05
This is a very practical feature: machine-core can dynamically generate pydantic-ai tools based on OpenAPI specifications. The workflow is as follows:
fetch_openapi_spec)ToolFilterManager to index and filter toolsrebuild_agent)This design allows the Agent to dynamically adjust the available tool set according to current task requirements, avoiding context bloat caused by loading too many irrelevant tools at once.
Section 06
machine-core has a built-in vector-based tool filtering mechanism:
ToolFilterManager uses vector similarity to index and filter toolsessential_tools to force inclusion of specific toolsThis mechanism significantly improves the Agent's tool selection efficiency, especially in scenarios with a large number of tools.
Section 07
The framework has built-in multiple file processing functions:
These capabilities enable the Agent to handle multimodal inputs, expanding application scenarios.
Section 08
Vector storage implementation based on LanceDB:
DocumentStore facade patternThis provides a solid infrastructure for RAG (Retrieval-Augmented Generation) applications.
Keep going with more reads from the same topic.
Splinter is a minimalist, high-performance key-value (KV) and vector storage system enabling zero-latency inter-process communication via shared memory and atomic operations. With only 766 lines of core code, it supports millions of operations per second and 768-dimensional vector storage, offering a new architectural approach for local LLM inference and data-intensive applications.
Folkering OS is the world's first AI-native bare-metal operating system, entirely written in Rust no_std without relying on Linux, POSIX, or libc. It can generate commands from scratch, compile them into WASM, and run them in 10 seconds, achieving true self-evolution.
This project explores how to use Large Language Models (LLMs) to detect logical vulnerabilities in Ethereum smart contracts, especially business logic flaws that are difficult for traditional static analysis tools to capture, such as economic manipulation and execution order errors.
This article details how to quickly deploy a virtual machine equipped with 4 NVIDIA V100 GPUs on Azure using Terraform to enable local inference of large language models with over 70B parameters. It covers a complete practical guide from infrastructure deployment, Ollama/vLLM dual-engine comparison testing, to cost optimization and actual performance data.