Neural Computer

Neural Computer is a local-first AI desktop app built with React + Tauri. It runs a local Node server for model orchestration/tools and a desktop shell for packaging.

Website: neural-computer.com

Note: the packaged app metadata is still named Neural Computer in src-tauri/tauri.conf.json and can be renamed separately.

What It Includes

• Tauri desktop app (src-tauri/)
• React frontend (*.tsx at repo root + components/)
• Local API/server runtime (server.mjs)
• Workspace + memory/skills runtime under workspace/ and services/

Requirements

• Node.js (current LTS recommended)
• npm
• Rust toolchain (required for Tauri dev/build)
• macOS/Linux/Windows system dependencies for Tauri v2
  See: Tauri prerequisites

Quick Start

1. Install dependencies:

   npm install

2. Start web client + local server (non-Tauri):

   npm run dev

   • Frontend: http://localhost:3000
   • Local API server: http://localhost:8787

3. Start the Tauri desktop app:

   npm run tauri:dev

Model Provider Setup

Provider/model selection is runtime-configurable in app Settings.

API keys can come from:

• In-app session credentials (preferred for local sessions)
• auth.json in the repo root or ~/.codex/auth.json (used for openai-codex token flow)
• Environment variables loaded via .env.local/.env

Common key names used in this codebase:

• GEMINI_API_KEY
• GOOGLE_API_KEY
• GOOGLE_GENERATIVE_AI_API_KEY
• GOOGLE_AI_API_KEY

Server port env override:

• NEURAL_COMPUTER_SERVER_PORT

Scripts

• npm run dev - Run frontend and server together
• npm run dev:client - Run Vite frontend only
• npm run dev:server - Run Node server only
• npm run tauri:dev - Run desktop app in development
• npm run build - Build frontend bundle
• npm run tauri:build - Build desktop bundles
• npm run typecheck - TypeScript checks
• npm run test - Run Vitest tests
• npm run validate - Typecheck + frontend build

Project Docs

• docs/onboarding-runtime.md
• docs/self-improvement-system.md
• docs/CONTINUOUS_USER_FIT_REPORT.md

Open Source Status

This repository is being prepared for public open source release. Current recommendation before public launch: add a top-level LICENSE file and contribution guidelines.