Skip to content

README.md

Latest commit

 

History

History
76 lines (62 loc) · 4.17 KB

README.md

File metadata and controls

76 lines (62 loc) · 4.17 KB

Scark

Scark is a desktop application scaffolded with Nextron (Electron + Next.js) that focuses on local/experimental AI workflows, web content extraction, and a fast developer-friendly UI. This repository provides the app shell and several integrated libraries to enable local inference, content processing, vector storage, and automated browsing — all packaged as an Electron desktop app.

NOTE: Scark is a work in progress. Core functionality is available, but many features are experimental and actively being developed. Expect new features, improvements, and breaking changes in upcoming updates.

Current capabilities

  • Electron + Nextron desktop app shell
    • Cross-platform desktop packaging via electron-builder.
  • Dynamic deliberation based thinking loop
    • agentLoop.js describes how the model deliberates actions
  • Standardized tool registry
    • Any tool initialized by the tool registry can be used by the model.
  • Local/experimental LLM support
    • Libraries present: @mlc-ai/web-llm and @xenova/transformers — used for running models locally (experimental).
  • Embeddings and vector store integration
    • chromadb and better-sqlite3 are included for storing and querying embeddings and doing vector similarity searches.
  • Web content extraction and document parsing
    • @mozilla/readability and jsdom are available for extracting readable article content from arbitrary web pages.
  • Automated, stealthy web interaction and scraping
    • Playwright (and playwright-extra) plus puppeteer-extra-plugin-stealth are included for headless/automated browsing and stealthy scraping when needed.
  • Markdown rendering and syntax highlighting
    • react-markdown and react-syntax-highlighter are present for rendering rich, readable content in the UI.
  • Persistent local settings
    • electron-store is used for saving app settings persistently across runs.
  • UI and animation tooling
    • @geist-ui/react, Tailwind CSS, framer-motion, gsap, lucide-react and next-themes are available for building a polished, responsive interface and animations.
  • Developer & testing tooling
    • Playwright config and tooling are included for end-to-end tests.
  • Packaging hooks
    • electron-builder config is present to help build distributable desktop installers.

Quick start

  1. Clone the repo
  2. Install dependencies
    • yarn or npm install
  3. Development
    • yarn dev (or npm run dev / pnpm run dev)
  4. Build for production
    • yarn build (or npm run build / pnpm run build)
  5. Packaging
    • electron-builder is configured in electron-builder.yml (postinstall hook in package.json installs app deps for building).

(See package.json for exact script names and current dev/build commands.)

Project layout (high level)

  • main/ — Electron main process code
  • renderer/ — Next.js / React renderer UI
  • services/, workers/, lib/ — service layer and auxiliary code (placeholders for current/next features)
  • resources/ — static app resources and assets
  • pipeline/ — automation/test pipelines and scripts

Roadmap / coming soon

  • Improved local model orchestration and easier model downloads for local inference (more model types & UI).
  • UI for building/searching vector stores and inspecting embeddings.
  • Workflow editor for automated scraping, content ingestion, and chunking pipelines.
  • Better error handling, auto-updates, and cross-platform packaging improvements.
  • User preferences, onboarding, and sample datasets to get started quickly.
  • Autonomous task execution capabilities within a sandbox.

If you have feature ideas, bug reports, or want to help implement features, please open an issue or submit a PR.

Contributing

Contributions are welcome. If you plan to contribute:

  • Open an issue describing the feature or bug first (so we can coordinate).
  • Follow the existing code style (Next + React + Tailwind patterns).

Notes & warnings

  • Many dependencies relate to experimental local ML usage. Running large models locally may require substantial disk space, memory, and compatible hardware.

Scark is evolving rapidly — this README reflects the repository's current state and intended direction. Expect more user-facing features and improvements soon.