DEV Community: ziuus

OmniChat Harness: Multi-Provider Agent Harness & Orchestration Framework with MCP

ziuus — Mon, 02 Nov 2026 10:50:44 +0000

Autonomous agents require robust execution harnesses that can route between multiple model providers (Anthropic Claude, OpenAI, Google Gemini, local Ollama/vLLM), handle real-time streaming deltas, and dynamically interact with Model Context Protocol (MCP) tool servers.

OmniChat Harness (GitHub: ziuus/omnichat-harness) is a TypeScript/Bun-powered orchestration framework and development environment built specifically for multi-provider agent development, streaming evaluation, and MCP server interop.

Key Architectural Capabilities

Unified Multi-Provider Gateway: Transparent abstraction over Claude 3.5 Sonnet, GPT-4o, Gemini 2.0 Flash, and local models.
Native Model Context Protocol (MCP): Full support for client connections to filesystem, database, browser, and custom MCP tools.
Real-Time Streaming Engine: Zero-lag Server-Sent Events (SSE) and WebSocket conduits for token streaming and structured tool call dispatch.
Agent Memory & Session Journaling: Append-only execution transcripts, prompt cache optimization, and replayable execution trees.

Quick Start with Bun

git clone https://github.com/ziuus/omnichat-harness.git
cd omnichat-harness
bun install
bun run dev

Explore the repository, star the project, and inspect our MCP integrations:
https://github.com/ziuus/omnichat-harness

Building a Sub-Millisecond AI Gateway in Go for Autonomous Coding Agents

ziuus — Mon, 02 Nov 2026 10:49:08 +0000

When building workflows with autonomous coding agents (like Claude Code, Cursor, or OpenCode), agent execution loops issue dozens of rapid LLM calls per minute. Standard Python-based proxies introduce 40ms–100ms of latency per hop, balloon to hundreds of megabytes in memory, and easily choke during parallel agent streaming.

To solve this, I built GoRoute (https://github.com/ziuus/goroute) — an ultra-lightweight, high-throughput AI gateway and dynamic LLM router written in Go.

The Latency & Reliability Problem

In autonomous coding agent sessions:

Compounding Proxy Latency: In a 30-step agent loop, 60ms of routing overhead adds 1.8 seconds of pure wait time.
Provider Rate Limits: Hitting a 429 on Anthropic or OpenAI kills the entire agent trajectory unless failovers are instantaneous.
Protocol Mismatches: Tool definitions and message schemas differ between OpenAI (function_call / tools) and Anthropic (messages API).

Architecture & Key Features

Sub-0.8ms Local Routing: Zero-allocation hot path, standard library HTTP + lightweight routing, baseline footprint <12MB RAM.
Universal Protocol Translation: Transparently translate OpenAI format to Anthropic messages format and vice-versa, allowing any coding tool to talk to any upstream model.
Dynamic Failover Cascades: If Claude 3.5 Sonnet hits rate-limits or 5xx errors, GoRoute cascades automatically to DeepSeek-V3 or GPT-4o with zero lost tokens.
Multi-Account Token Pooling: Distribute requests round-robin across multiple API keys to maximize quota without manual key rotation.
Native MCP Compatibility: Designed to sit cleanly alongside Model Context Protocol servers.

Quick Start

Run locally via Go:

go install github.com/ziuus/goroute@latest
goroute --port 8080 --config config.yaml

Or via Docker:

docker run -p 8080:8080 -v $(pwd)/config.yaml:/app/config.yaml ghcr.io/ziuus/goroute:latest

Point your agent or SDK to http://localhost:8080/v1 and you are ready.

Open Source

GoRoute is 100% open source under the MIT License.

GitHub Repository: https://github.com/ziuus/goroute
Live Demo & Docs: https://goroute-website.vercel.app

Feedback, issue reports, and benchmarks are warmly welcomed! If you are building agentic tools or multi-model pipelines, give it a try.h

Liquid Glass: Hyper-Fluid Physics & Liquid Glass Rendering Engine for Linux and Wayland

ziuus — Mon, 02 Nov 2026 10:39:32 +0000

Modern desktop user experiences on macOS and mobile platforms have long set a high bar with dynamic fluid glassmorphism, depth shaders, and physics-based inertia. On Linux and Wayland, desktop environments are now entering an era of rich GPU-accelerated compositing.

Liquid Glass (GitHub: ziuus/liquid-glass) is a lightweight, high-performance rendering engine and physics library written in pure Rust, designed to bring real-time liquid glass, chromatic refraction, and fluid spring physics to Linux and Wayland window managers (such as Hyprland, Sway, and custom Smithay-based compositors).

Key Capabilities & Engine Features

Real-Time Chromatic Dispersion & Refraction: Sub-surface scattering and physically approximated light bending through frosted and curved liquid surfaces.
Spring-Damper Fluid Inertia: Natural spring physics for gestures, window borders, floating docks, and interactive widgets.
Zero-Copy Wayland Buffer Passing: Low overhead surface rendering utilizing Vulkan and OpenGL ES compute passes directly onto Wayland subsurfaces.
Modular Architecture:
- glass-core: Mathematical physics model, spatial geometry, and spring state solvers.
- glass-platform: Wayland wlroots and XDG shell surface attachments.
- glass-ui: Reusable UI primitives with fluid glass styling.

Building & Running Examples

Clone the repository and run the interactive demo:

git clone https://github.com/ziuus/liquid-glass.git
cd liquid-glass
cargo run --example fluid_demo --release

Explore the project, star the repo, and join the discussion:
https://github.com/ziuus/liquid-glass

Settings TUI: A Unified Native Control Center for Linux Power Users

ziuus — Mon, 02 Nov 2026 10:39:03 +0000

Modern Linux workflows and tiling window managers (Hyprland, Sway, i3) offer unparalleled customization, but configuring hardware devices often requires memorizing and juggling incompatible command-line tools: nmcli for Wi-Fi, bluetoothctl for Bluetooth pairings, pactl / wpctl for PipeWire audio streams, and systemctl for systemd services.

Settings TUI (GitHub: ziuus/settings-tui) brings all of these hardware and subsystem controls into a fast, keyboard-driven native terminal console written in Rust.

Core Architecture & Capabilities

Instant Startup (<10ms): Built in Rust with minimal memory overhead (<15MB RSS).
Direct System D-Bus Integration: Communicates directly over system and user D-Bus sockets rather than parsing unbuffered shell output.
PipeWire & WirePlumber Support: Live volume level metering, output device routing, and per-application audio stream volume balancing.
NetworkManager & IWD: Wi-Fi network scanning, connection authentication, static IP setup, and VPN profile toggling.
systemd Unit Manager: Inspect unit state, start/stop/restart daemons, and view live journal logs with ANSI highlighting.
Hyprland & Display Configuration: Fast monitor resolution, refresh rate, and scaling adjustments.

Installation & Getting Started

Install via Cargo:

cargo install settings-tui

Or build from source:

git clone https://github.com/ziuus/settings-tui.git
cd settings-tui
cargo build --release
./target/release/settings-tui

Check out the repository and contribute on GitHub:
https://github.com/ziuus/settings-tui

Orbit: A High-Performance Terminal Workspace & System Monitor in Rust with WASM Plugins

ziuus — Mon, 02 Nov 2026 10:21:19 +0000

Most terminal dashboards force a compromise: either they are fast but rigid (compiled C/Rust with fixed layouts and hardcoded charts), or they are extensible but slow (heavy Python or Electron apps).

To bridge this gap, I built Orbit (https://github.com/ziuus/orbit-tui) — an extensible, keyboard-driven developer dashboard and real-time system monitor written in Rust.

What Makes Orbit Different?

WASM Plugin Architecture: Write custom telemetry plugins or workspace integrations in any language that compiles to WebAssembly. Orbit dynamically loads them at runtime without recompiling the core binary.
Sub-millisecond Input Latency: Built with Ratatui and Crossterm using decoupled async background data pollers. UI rendering is locked at a steady 60 FPS while background workers stream stats.
Modal Vim Navigation: Everything is keyboard-controllable. Switch panels with h/j/k/l, inspect container logs with <Enter>, and customize keymaps directly in YAML.
All-in-One Developer Views:
- Host metrics: Per-core CPU utilization, memory pressure, NVMe I/O, network socket throughput.
- Container orchestration: Real-time Docker container states and log streams.
- Process tree with instant fuzzy filtering.

Installation & Quick Start

Via Cargo:

cargo install orbit-tui
orbit

Or download pre-compiled releases from GitHub: https://github.com/ziuus/orbit-tui/releases

Open Source & Community

Orbit is 100% open-source under the MIT license.

GitHub Repository: https://github.com/ziuus/orbit-tui
Website & Documentation: https://orbit-website.vercel.app

Stars, issues, and custom WASM widget ideas are welcome! Give it a spin in your terminal and let me know your thoughts.

Building a Sub-Millisecond AI Gateway in Go for Autonomous Coding Agents

ziuus — Mon, 02 Nov 2026 10:20:50 +0000

To solve this, I built GoRoute (https://github.com/ziuus/goroute) — an ultra-lightweight, high-throughput AI gateway and dynamic LLM router written in Go.

Key Features

Sub-0.8ms Local Routing: Baseline footprint <12MB RAM.
Universal Protocol Translation: Translate OpenAI format to Anthropic messages format.
Dynamic Failover Cascades: Automatic fallback from Claude 3.5 Sonnet to DeepSeek-V3 or GPT-4o.
Multi-Account Token Pooling: Distribute requests round-robin across multiple API keys.

GitHub: https://github.com/ziuus/goroute
Demo: https://goroute-website.vercel.app