Toolpack SDK ships with built-in support for OpenAI, Anthropic, Gemini, and Ollama. But sometimes you need to integrate a different provider -- xAI's Grok, Groq, DeepSeek, Azure OpenAI, or your company's own model. Custom providers make that possible without giving up any SDK features like tool calling, streaming, or embeddings.

This guide walks through exactly how to build one, based on the real xAI adapter used in the Toolpack CLI project.

Configuration is the backbone of any flexible SDK. Toolpack SDK implements a sophisticated hierarchical configuration system that balances simplicity for quick starts with power for complex production deployments. Whether you're integrating AI into your app or managing multiple projects, understanding how configuration works will unlock the full potential of the SDK.

Let's dive deep into how toolpack.config.json works, the three-tier hierarchy, and best practices for managing configurations across different environments.

AI agents need to access real-time information from the web. Whether it's searching for current events, fetching documentation, or scraping data, Toolpack SDK's web tools make it seamless.

In this post, we'll explore the web search fallback system (Tavily → Brave → DuckDuckGo Lite) and the web fetch tools that power web-aware AI applications.

Tools are the foundation of what makes AI agents useful. While LLMs are great at reasoning and generating text, they need tools to interact with the real world—reading files, making API calls, querying databases, and executing code.

Toolpack SDK ships with a deep catalog of built-in tools (with more landing every release) and makes it trivial to add your own. Let's explore how the tool system works and why it matters.

The AI tooling ecosystem is growing fast—but it's also fragmenting. Developers face vendor lock-in, inconsistent tool execution, and a lack of standardization when building AI-powered systems. Toolpack SDK is our answer to these problems.