incognide  by NPC-Worldwide

AI-powered unified workspace for development and productivity

Created 1 year ago
255 stars

Top 98.8% on SourcePulse

GitHubView on GitHub
Project Summary

Incognide unifies chat, code, documents, web browsing, and media into a single, tileable workspace, aiming to eliminate distractions and context switching for developers, researchers, and power users. It offers a structured, auto-saving environment with composable automations and AI agent capabilities, allowing users to seamlessly manage complex projects and workflows.

How It Works

Incognide employs an Electron + React frontend with a Python Flask backend (npcpy). Its core design features a path-based organization and auto-saving workspaces. It integrates diverse tools like code editors, PDF viewers/annotators, web browsers, and terminals, all within a composable, tileable interface. A key differentiator is its AI agent system, which can interact with these tools, execute workflows (Jinxs) combining natural language and code, and manage context. Users can configure local or cloud-based AI models and fine-tune models directly within the application.

Quick Start & Requirements

  • Primary install / run command: Download the installer for Linux, macOS, or Windows from enpisi.com/incognide.
  • Non-default prerequisites and dependencies: For local models: Ollama, LM Studio, or a llama.cpp server. For cloud models: API keys for OpenAI, Anthropic, Gemini, etc. For development setup: Node.js 16+ and npm.
  • Links: Installer: enpisi.com/incognide. Community: Discord (link not provided in README, but implied). Issues/Discussions: GitHub Issues/Discussions.

Highlighted Details

  • Unified workspace integrating chat, code editing, document management (DOCX, XLSX, PPTX, PDF), web browsing, and terminals.
  • AI agents capable of tool use, including browser automation, file operations, and terminal command execution.
  • Built-in support for fine-tuning custom image (LoRA) and text models using curated data from conversations and memories, with a focus on local training.
  • Composable automation workflows (Jinxs) that chain natural language prompts with code execution, including database querying and agent integration.
  • Cloud sync with end-to-end encryption for workspace state, conversations, and settings across multiple devices.

Maintenance & Community

The project maintains an active community via GitHub Issues and Discussions. It is part of a broader "NPC Ecosystem" including npcpy, npcsh, and npcts. A Discord community link is implied but not directly provided in the README.

Licensing & Compatibility

Incognide is licensed under AGPLv3 with additional terms that prohibit third-party SaaS services and packaged resale. The AGPLv3 license is copyleft, meaning derivative works must also be open-sourced under the same license. The additional terms may restrict commercial use, particularly for service providers or those intending to bundle the software.

Limitations & Caveats

The AGPLv3 license and its additional terms impose significant restrictions on commercial use and distribution, potentially limiting adoption in closed-source or SaaS environments. A "Demo Video Updated version coming soon" note suggests the project may be under active, potentially breaking, development. macOS users require specific permissions for voice input and screenshot features.

Health Check
Last Commit

21 hours ago

Responsiveness

Inactive

Pull Requests (30d)
20
Issues (30d)
55
Star History
23 stars in the last 30 days

Explore Similar Projects

Feedback? Help us improve.