Show HN: Deepdrone – Controls drones with natural language through LLMs (github.com)

Deepdrone has emerged as an innovative tool that empowers users to control drones using natural language commands through large language models (LLMs). This open-source project supports various AI providers, including OpenAI, Anthropic, and Google, while also offering a local option with Ollama. Users can easily set up the system by installing the necessary dependencies and launching a web interface, allowing for conversational interactions such as "Take off to 20 meters" and "Return home and land safely." Additionally, Deepdrone integrates DroneKit for real flight control and includes a simulator feature to test commands without hardware. The significance of Deepdrone lies in its potential to democratize drone operations, making advanced control accessible to non-experts through natural language processing. The project introduces real-time telemetry features, enabling users to monitor drone status while executing complex maneuvers. With its support for low-latency UDP control for Webots C-based simulators, Deepdrone offers efficient communication free from MAVLink overhead. This not only enhances the responsiveness of drone commands but also sets a new standard for interactive drone control, paving the way for a broader application of AI in unmanned aerial vehicle operations.