Magnitude

# MISSION: Scaffold "Magnitude" Production Architecture
You are an expert Senior Staff Software Architect and Full-Stack Engineer. Your mission is to scaffold and implement a production-grade, highly reliable, and modular codebase following the proven architecture of **Magnitude**.
---
## 1. PROJECT SPECIFICATIONS & BENCHMARK
- **Reference Architecture**: Magnitude
- **What It Does**: Magnitude is an open-source local inference engine and desktop app. It profiles the machine, recommends and tunes open models for that hardware, runs them through a Rust llama.cpp-based server, and connects them to coding agents in one click.
- **Domain & Category**: Local LLM Inference Engine & Hardware Profiler
- **Production Scale**: 2-5 people
- **Development Mode**: HYBRID
- **Architectural Rationale**: Performance-critical inference sits in a Rust workspace around vendored llama.cpp, while the app, CLI, agent and model management live in a Bun + Effect TypeScript monorepo. The native engine can be pinned and benchmarked on its own while the product layer changes quickly.
- **Live Website Reference**: https://magnitude.dev
- **Source Repository**: https://github.com/magnitudedev/magnitude
---
## 2. PRODUCTION TECH STACK
- **Full Stack Array**: TypeScript, Rust, Bun, Effect, llama.cpp, Axum, Electron, React, Vite, Tailwind CSS, Turborepo, OpenTelemetry
- **Primary Language(s)**: TypeScript, Rust, C, Svelte, Python
- **License of the reference repo**: Apache-2.0
- **Frontend**: React — React 19 UI in web/ (Base UI, shadcn, Monaco, cmdk), shared by the desktop app; Electron — Cross-platform desktop shell built with electron-vite that bundles the CLI and daemon; Vite — Bundler for the web UI and, through electron-vite, the desktop renderer; Tailwind CSS — Styling for the web and desktop UIs (Tailwind 4 via @tailwindcss/vite)
- **Backend & APIs**: Rust — Inference workspace (inference/crates/icn-*): engine, hardware profiling, catalog, server; llama.cpp — Model execution through llama-cpp-2 bindings vendored as a pinned native submodule; Axum — HTTP server for the local inference API in the icn-server crate; Bun — Runtime and package manager for the TypeScript workspaces, CLI and scripts; Effect — Typed effect system used across CLI, SDK, agent and ICN packages, including @effect/rpc
- **Infrastructure & deploy**: Turborepo — Orchestrates build and typecheck across the Bun workspaces; OpenTelemetry — OTLP traces and logs from the Rust inference server; GitHub Actions — Release, desktop-native, integration and update-acceptance workflows per platform
- **Tooling, testing & ops**: Harness connections — packages/harness-connections wires local models into agents such as Pi, OpenCode and Hermes; Hugging Face Hub — Model downloads via the hf-hub crate for the catalog and model management
---
## 3. KEY ARCHITECTURAL DECISIONS (audited from https://github.com/magnitudedev/magnitude @ fa33238)
1. **TypeScript product layer on top of a separate Rust inference workspace**: The root Bun workspace (cli, desktop, web, packages/*) holds the app, agent and client logic, while inference/ is its own Cargo workspace (icn-engine, icn-hardware, icn-speculative, icn-server, …) exposed to TypeScript through packages/icn and packages/icn-protocol.
2. **llama.cpp vendored as a pinned native dependency**: inference/Cargo.toml depends on llama-cpp-2 from native/llama-cpp-rs (a git submodule, see .gitmodules) and inference/native-pin.toml plus verify:native-pin and verify:safe-bindings scripts guard the exact native revision and the unsafe boundary.
3. **Hardware profiling before model choice**: The icn-hardware crate (with an icn-model-assessment binary) and inference/catalog back the app's claim of estimating tokens per second for every catalog model before download; parity and benchmark crates check results against references.
4. **Effect as the application backbone**: CLI, desktop, web, agent and ICN packages all depend on effect 3 with @effect/platform and @effect/rpc, and patches/ carries a local fix for @effect-atom/atom, so services, errors and RPC share one typed runtime model.
5. **One desktop app ships the CLI and daemon**: desktop/ (Electron, electron-vite) runs build:native from packages/daemon-management before bundling, so installing the app also installs the magnitude CLI and the local daemon agents connect to.
---
## 4. NON-NEGOTIABLE ARCHITECTURAL GUARDRAILS
1. **Monorepo & Modular Separation**:
   - Structure as a Turborepo monorepo with strict package boundaries:
     - `apps/web`: Application UI, routing, layouts, and server endpoints.
     - `packages/ui`: Shared design tokens, CSS variables, and Radix UI primitive components.
     - `packages/db`: Database schemas, client singleton, declarative migrations, and seed scripts.
     - `packages/config`: Shared TypeScript, ESLint, and build configurations.
2. **Strict Type Safety & Zero `any` Policy**:
   - Enable `strict: true`, `noImplicitAny: true`, and `strictNullChecks: true`.
   - Validate ALL external inputs, API request bodies, and query parameters with **Zod** schemas before execution.
3. **Frontend & Rendering Guidelines**:
   - Utilize Vite 6 with TanStack Router for fully type-safe routing. Manage server state and caching via TanStack Query v5 with optimistic UI updates.
4. **Design System & Aesthetics**:
   - Keep every color, radius, shadow and font in a single token file (CSS variables) and consume tokens everywhere; never hardcode hex values in components.
   - Prefer crisp 1px borders and one subtle shadow scale over blurry default shadows. Pair one sans-serif for body/headings with one monospace for tags, badges, metrics, and code.
5. **Data Layer & Reliability**:
   - Write declarative schema definitions with foreign keys, composite indexes on queried filters, and automated timestamp triggers.
   - Use connection pooling and prepared statements for serverless database execution.
---
## 5. STEP-BY-STEP SCAFFOLDING ROADMAP
- **Phase 1: Workspace & Root Config**: Initialize package manager, monorepo configuration (`turbo.json`, `tsconfig.base.json`, `package.json`).
- **Phase 2: Database Schema & Client**: Set up the data layer: client, connection pool, models, and migration scripts.
- **Phase 3: Design Tokens & UI Primitives**: Build accessible `Button`, `Input`, `Card`, `Badge`, and layout wrappers inside `packages/ui`.
- **Phase 4: Core Application Routes & Handlers**: Implement primary authentication, user session handling, and application routes.
- **Phase 5: Quality Assurance & Build Verification**: Run `tsc --noEmit`, ESLint, Prettier, and smoke test suites to ensure zero compilation or runtime errors.
---
## 6. EXECUTION INSTRUCTIONS
1. Review all specifications, architectural guardrails, and stack choices above.
2. Present the full monorepo directory tree structure.
3. Systematically generate the complete, production-ready codebase according to the 5-phase roadmap above — starting with the root workspace setup, followed by the database schema, UI design system package, and full-stack application routes until the repository is fully scaffolded and ready to run.