CrossKlaw

CrossKlaw is a powerful AI platform that works as a single, self-contained program. It's designed to be secure, work even without an internet connection, and understand the hardware it's running on. This makes it flexible enough to run on devices like a Raspberry Pi.

Benefits

CrossKlaw offers many advantages. Its core is a simple Go program that doesn't need anything else to run, making it easy to set up and use. It can connect to a wide variety of communication tools like Telegram, Discord, and WhatsApp, and it works with many different AI language models from companies like OpenAI and Google. It can intelligently choose the best AI model for a task and switch to another if one isn't working. CrossKlaw can also connect with other AI agents, control smart devices, and run custom code safely within a protected environment. It includes features for voice commands, requires human approval for important actions, and provides detailed insights into its performance and costs. Security is a priority, with defenses against prompt attacks and encrypted settings.

Use Cases

This platform is useful for many applications. It can act as a central hub for AI interactions across different messaging apps. Businesses can use it to build custom AI assistants that connect to their existing systems and control IoT devices. Developers can integrate it into applications that require advanced AI capabilities, like analyzing documents or managing complex workflows. Its ability to run on small devices makes it suitable for edge computing scenarios. The multi-agent features allow for complex AI systems where different agents collaborate on tasks.

Pricing (ONLY include if available)

CrossKlaw is available for personal, educational, and evaluation use. Commercial use requires a license from Black Stick Ltd.

Vibes (ONLY include if available)

Additional Information (ONLY include if available)

CrossKlaw is built as a single Go binary, approximately 50MB in size, with no external dependencies. It supports 14 communication channels and integrates with 11 LLM providers. It features an Agent-to-Agent (A2A) protocol server and acts as a Model Context Protocol (MCP) client and server. It uses WebAssembly for sandboxing skills and includes features for voice input and output, human-in-the-loop approval, and OpenTelemetry tracing for monitoring. It also has an IoT bridge for controlling devices and supports document retrieval augmented generation (RAG) with persistent memory stored in SQLite. The platform comes with 19 bundled skills and 10 sector templates. It includes a cron-style job scheduler. The gateway API defaults to port 18800, and the WebChat UI uses port 18801.