SUZENT / Docs
THE SUZENT HANDBOOK

Architecture

How Suzent's runtime, tools, memory, and clients work today.

Agent Communication

Suzent treats every agent-backed conversation as an addressable session. The stable agent_id is its chat ID; kind describes how the session originated: interactive , subagent , cro

Canvas (A2UI)

The Canvas is a structured UI layer that lets the agent render interactive components — tables, forms, cards, buttons, badges — directly in the Suzent interface, alongside the chat

Memory Architecture

How memory is written, consolidated, indexed, and read. Companions: Memory for how the dream behaves from the outside, Memory internals for the class-by-class reference. Three tier

Memory Internals

This page describes the implementation-level structure of Suzent memory as it exists today. Layers Layer Responsibility --- --- Agent prompt and tools memory_search , block_update

Native mobile clients

Architecture decision Suzent mobile uses SwiftUI on iOS and Kotlin/Jetpack Compose on Android. Both applications live in this monorepo; Python remains managed by uv, Android by Gra

Model capabilities and roles

This page is for contributors. For choosing providers and assigning model roles as a user, see Models & Providers. Suzent keeps a registry of per-model metadata — context window, c

Chat Post-Processing

When an agent turn finishes, Suzent runs a series of background tasks — writing the transcript, updating memory, compressing context, and saving state. This page explains how that

Stream recovery protocol

Desktop chat sends, retry/edit, steering, approval resume, heartbeat and Canvas turns can recover their in-progress output after a page reload or connection loss. Native and ACP pr

System Reminders & File Citations

System reminders carry runtime-authored context for the model. File citations render source references for the user. Both use Private Use Area (PUA) markers, but invisible formatti

Tool system

This page is for contributors. For what tools do from a user's point of view, see Tools. Tools extend the agent's capabilities, allowing it to interact with the filesystem, web, me