Developer Guide¶

This section is meant for developers who want to use pybrid within their own projects and for contributors to anabrid's software stack. pybrid has two main functions: constructing, serializing and deserializing the entity object model for analog devices, and serving as the runtime layer that handles communication, access control, and connection management for connected analog devices. The developer docs split along those two lines into the entity object representation on the one hand, and the data and messaging protocol together with the native networking code on the other. Coming from the architecture overview, especially from the LUCIstack perspective, we recommend starting with the entity object model.

On AI and agentic usage¶

As software developers and researchers, we live in wild times, with AI agents becoming better every month. The broad access to frontier-level intelligence also changes the value proposition of documentation like this: instead of painstakingly including code and adding structured code documentation, down to docstrings, it has become the norm to convey intent, architecture and core guidelines in the documentation. Code itself is closer to a "living organism", changing frequently and quickly leading to out-of-sync documentation.

Following this idiom, this section of the documentation offers enough information and overview to convey the guidelines and concepts governing the development of pybrid, but leaves the links to the actual code and the instructions on how to use it to agents. With the start of 2026, agents have become really good at search and synthesis and are now mature solutions for Q & A with middle-sized codebases such as this. We recommend pointing your agent at the docs/ folder first, especially the developer documentation, or bootstrapping it from the AGENTS.md file (rename to CLAUDE.md if you are a Claude Code user).