How We Build Axonris

This is not Forge's own architecture — Forge, the app you download, breaks your request into planning, coding, and review steps with a bounded tool loop that can read your files and check its own changes before writing them, running on your machine (local Ollama, with a BYOK cloud fallback if local is unreachable). This diagram is our own internal engineering process: how a coordinated team of AI agents actually builds and maintains Axonris itself, on our dev machines, before anything ships to you.

Internal engineering diagram: how the Axonris team's own development requests move through a Swarm Manager to specialized agents (Claude, Gemini, DeepSeek, local LLM), then through a Safe Sandbox before a change ships — this is our build process, not Forge's own runtime architecture