How to create your own custom conversation app on Reachy Mini 🤖💬

2026-02-11 · Source: HuggingFace · Field: Technology & Digital — Robotics & Autonomous Systems, Artificial Intelligence & Machine Learning, Software Development & Engineering · Depth: Intermediate, medium

Summary

This content details the process of creating and customizing applications for the Reachy Mini, an open-source robot developed by Pollen Robotics and Hugging Face. Reachy Mini apps, written in JS or Python, are distributed via Hugging Face Spaces, functioning as a community app store. The Reachy Mini SDK, available on GitHub, provides tools like the Reachy Mini App Assistant for development, testing, and publishing, and supports simulation for those without a physical robot. The core example is the "conversation app," which uses the OpenAI real-time API for interactive dialogue. Users can customize this app's prompt and tools via a web interface or by forking the app to add custom tools, as demonstrated with an "astronomy buddy" app that enables the robot to identify and physically point to celestial objects. Custom tools involve subclassing a tool class, defining its name, description, parameter schema, and a `call` method, with dependencies managed via `pyproject.toml` and `uv`. The process involves modifying `instruction.txt` for persona, `tools.txt` for tool listing, and adding Python files for custom tool logic. Local testing requires an OpenAI API key and the `gradual` flag, while publishing is done via the App Assistant.

Key takeaway

For AI Engineers developing interactive robot applications, the Reachy Mini SDK and its app development workflow offer a streamlined path. You can rapidly prototype and deploy custom conversational agents by leveraging the existing conversation app, extending its capabilities with custom tools, and defining unique personas. This approach significantly reduces development overhead, allowing you to focus on specific functionalities like integrating external data or physical robot actions, rather than building the entire interaction pipeline from scratch.

Key insights

Reachy Mini's open platform allows extensive app customization and distribution through Hugging Face Spaces.

Principles

• Open-source robotics fosters community-driven app development.
• Tool calling extends LLM capabilities with custom functions.
• Modular design simplifies app customization and extension.

Method

To create a custom Reachy Mini app, fork an existing app (e.g., conversation app), define the robot's persona in `instruction.txt`, list available tools in `tools.txt`, and implement custom tool logic as Python files within the profile folder.

In practice

• Use Reachy Mini App Assistant to create, check, and publish apps.
• Customize LLM behavior by modifying system prompts and exposing tools.
• Develop custom tools by subclassing `Tool` and defining `call` methods.

Topics

• Reachy Mini Robot
• Robot SDK
• Hugging Face Spaces
• LLM Tool Calling
• Conversational AI

Best for: Robotics Engineer, AI Engineer, Software Engineer

Related on AIssential

• Reachy Mini at Nvidia's Jensen CES keynote — Open-source advancements enable complex, voice-controlled robotic interactions with ease.
• Adding MCP Tools to Reachy Mini — Reachy Mini now integrates remote tools from Hugging Face Spaces via MCP, expanding capabilities without local code modifications.
• Reachy Mini: the $300 open source robot you can actually hack — Andres Marafioti, Hugging Face — Democratizing robot interaction requires affordable, hackable hardware and robust open-source voice AI tools.
• What Is Conversational AI Development? A Practical 2026 Guide — Modern conversational AI systems integrate NLU, LLMs, and dialogue management to perform complex, context-aware, multi-step tasks.
• NVIDIA brings agents to life with DGX Spark and Reachy Mini — NVIDIA's DGX Spark and Reachy Mini enable building private, customizable, real-world AI agents using open models.
• How to build your own AI developer tools with Claude Code | CJ Hess (Tenex) — Custom AI-native tools enhance developer workflows by transforming AI-generated plans into interactive visual diagrams.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by HuggingFace.