ChromatUI

ChromaTUI is a kernel-level terminal UI foundation for AI development workloads. It prioritizes deterministic rendering, measurable runtime behavior, and audit-friendly decision systems over ad-hoc widget magic.

What You Get

• Deterministic runtime loop with explicit events, commands, and effects
• Diff-based renderer with sparse update optimization and ANSI presentation
• Inline and alt-screen output modes with synchronized update bracketing
• Bayesian/conformal/CUSUM-driven runtime controls for strategy and risk decisions
• Widget and extras layers for application composition

Design Principles

• Correctness over cleverness: predictable terminal state is non-negotiable
• Deterministic output: explicit frame -> buffer -> diff -> presenter pipeline
• Inline-first ergonomics: preserve scrollback while keeping UI chrome stable
• Layered architecture: core -> render -> runtime -> widgets, with clear boundaries
• Evidence before claims: conformance and validation are tracked in-repo

Workspace Crates

Crate	Purpose
chromatui	Public facade and re-exports
chromatui-core	Terminal lifecycle, events, capabilities, writer
chromatui-render	Content/buffer diffing, regions, ANSI presenter
chromatui-runtime	Deterministic event loop and frame pipeline
chromatui-layout	Flex and grid layout primitives
chromatui-style	Colors, themes, style escape generation
chromatui-text	Text spans, rope structures
chromatui-widgets	Focus, input, list/table, palette, modal utilities
chromatui-extras	Optional add-ons (including deterministic visual FX screen)
chromatui-harness	Snapshot and harness support
chromatui-pty	PTY utilities for terminal testing
chromatui-simd	Safe performance helpers
chromatui-algorithms	Bayesian/diff/control/statistical algorithm library

Architecture

Input Layer -> Runtime Loop -> Render Kernel -> Output Layer

TerminalSession (core)
  -> Event stream
  -> update(Event) => (Model', Cmd)
  -> view(Model) => Frame
  -> Frame -> Buffer/Content -> BufferDiff
  -> Presenter -> ANSI bytes
  -> TerminalWriter flush

High-level stack:

Application
  Widgets
    Layout
      Runtime
        Render
          Core

For full equations and algorithm catalog, see docs/architecture.md. For implementation evidence and verification traces, see docs/conformance-matrix.md.

Notable Runtime/Math Systems

• Bayesian command scoring and evidence ledgers
• Bayesian diff strategy selection with posterior change-rate tracking
• BOCPD resize regime detection and Bayes-factor coalescing evidence
• Conformal risk gating and residual alerting
• E-process + adaptive betting (GRAPA-style updates)
• CUSUM drift detection and hover stabilization
• VOI-driven expensive sampling decisions
• PI/PID pacing control with MPC objective evaluation support
• Count-Min sketch + W-TinyLFU admission primitives and PAC-Bayes bound helper

Quick Start

Prerequisites:

• Rust stable toolchain
• Cargo

Build workspace:

cargo check

Run tests:

cargo test

Run visual validation example:

cargo run -p chromatui-runtime --example visual_validation

Run render benchmarks:

cargo bench -p chromatui-render --bench diff_bench -- --sample-size 10

API Entry Point

Use the facade crate in crates/chromatui/src/lib.rs for re-exported runtime/core/render primitives:

use chromatui::{DeterministicRuntime, DiffRenderer, Event, OutputMode, TerminalWriter};

Then build your model loop around Event -> update -> view and render through the runtime pipeline.

Documentation Map

• docs/architecture.md: full architecture + algorithm reference
• docs/conformance-matrix.md: PASS/PARTIAL/MISSING matrix and verification evidence
• docs/plans/2026-02-22-kernel-tui-foundation-design.md: foundation design plan
• docs/plans/2026-02-22-kernel-tui-foundation.md: implementation plan

Community and Project Health

• Contributing guide: CONTRIBUTING.md
• Code of conduct: CODE_OF_CONDUCT.md
• Security policy: SECURITY.md
• Support channels: SUPPORT.md
• Changelog: CHANGELOG.md

License

MIT License with OpenAI/Anthropic rider.

• Primary license: LICENSE
• Rider: LICENSE-OPENAI-ANTHROPIC-RIDER.md

Copyright (c) 2026 Mike Holownych