Hi HN, I'm Pooria, the author of Intellistant. I built Intellistant because I was frustrated with the performance overhead of Python-based multi-agent frameworks like LangChain and CrewAI when running complex, long-running agent workflows with local LLMs. Even on high-end hardware, Python solutions often felt sluggish, especially during tool execution and agent coordination. Intellistant is a from-scratch, production-ready multi-agent framework written in modern C++23. Key highlights:

- 10–50× faster than equivalent Python implementations (measured on real software development tasks) - Full support for the Model Context Protocol (MCP) – works seamlessly with Claude, local models via Ollama/LM Studio, and more - Specialized agents for software engineering: Planner, Architect, Coder, Tester, Debugger, Reviewer - Zero runtime dependencies beyond the LLM backend – compiles to a single static binary - Docker one-liner for instant startup, plus REST API and CLI - 100% test coverage and CI/CD

You can try it in under a minute: docker run -it --rm -e OLLAMA_BASE_URL=http://host.docker.internal:11434 ghcr.io/pooriayousefi/intellistant:latest

(or build from source with CMake). The project is still young but already stable enough for real workloads. Roadmap includes Web UI, persistent memory, and more built-in tools. I'd love your feedback – especially if you try running an agent task, hit any issues, or have ideas for improvements. Happy to help with setup or answer questions! Thanks for checking it out.