Joanium

agent
Guvenlik Denetimi
Uyari
Health Uyari
  • License — License: MIT
  • Description — Repository has a description
  • Active repo — Last push 0 days ago
  • Low visibility — Only 5 GitHub stars
Code Uyari
  • fs module — File system access in .github/workflows/upload-release.yml
Permissions Gecti
  • Permissions — No dangerous permissions requested
Purpose
Joanium is a feature-rich desktop AI workspace built with Electron that integrates multi-model chat, local project tooling, automations, and AI agents into a single interface.

Security Assessment
Overall risk: Medium. The application is local-first by design, storing your chats and runtime state in a local `Data/` folder. However, because it is a project-aware workspace, it inherently interacts with local file systems, git, and terminal environments. The audit caught file system access within a GitHub Actions workflow (`upload-release.yml`), which is standard for publishing releases and not malicious. No hardcoded secrets were found and no dangerous permissions are requested. The biggest consideration is its extensive feature set: an app with deep terminal, file, and external connector access (like Gmail and GitHub) requires a high degree of trust, as it handles highly sensitive data.

Quality Assessment
The project uses the permissive MIT license and has a clear description. Development appears highly active, with the last repository push occurring today. However, community trust and visibility are currently very low. With only 5 GitHub stars, the tool has not yet been broadly adopted or thoroughly vetted by the open-source community.

Verdict
Use with caution. The project is active and licensed properly, but its low community visibility combined with deep local system and external account access means you should audit the code yourself before integrating sensitive workflows.
SUMMARY

Your smart, reliable, and friendly personal AI assistant.

README.md
Screenshot 2026-03-29 204551

Your desktop AI workspace. Local-first, project-aware, and built to actually help you get work done. ⚡

Website: Joanium.com

Joanium is a desktop AI app built with Electron for people who want more than a chat box in a shiny wrapper. It brings together multi-model chat, real project folders, automations, AI agents, MCP servers, connectors, skills, personas, and local data storage in one focused workspace.

It is serious software with range. Fast when you need speed, flexible when your workflow gets weird, and local-first by default so your machine still feels like yours.

What Joanium Actually Gives You

  • 🤖 Multi-model chat that does more than answer prompts.
  • 📁 Project-aware tooling with files, terminal, git, and workspace context.
  • ⏱️ Automations for repeatable tasks you do not want to babysit.
  • 🧠 AI agents for monitoring, summaries, triage, and recurring analysis.
  • 🔌 Live connectors like Gmail and GitHub that bring real context into the app.
  • 🛠️ MCP servers, skills, and personas so the app can adapt to how you work.
  • 🔒 Local-first storage under Data/, so your chats and runtime state stay on your machine.

Why It Hits Different

  • Not another disposable AI tab
    Joanium is designed like a workspace, not a demo.

  • Built for real workflows
    Chat connects to projects, tools, automations, agents, and external context instead of living in isolation.

  • Power without turning into chaos
    The architecture stays modular, so the app can grow without becoming a spaghetti pile.

  • Personality without losing credibility
    It is flexible, expressive, and useful without feeling gimmicky.

Core Areas

  • 💬 Chat - multi-model conversations with tools, connectors, and MCP support.
  • 🗂️ Projects - workspace-aware context with project-scoped chats and tooling.
  • ⚙️ Automations - scheduled action chains for deterministic work.
  • 🧠 Agents - scheduled AI jobs that collect data, reason over it, and produce outputs.
  • 📡 Events - a live operational timeline for runs, failures, skips, and active jobs.
  • 📚 Skills - local Markdown-based behavior packs that shape how the assistant works.
  • 🎭 Personas - switch the assistant's identity, tone, and framing.
  • 📈 Usage - local visibility into model activity and cost.

Quick Start

# Prerequisites: Node.js 18+ and npm

git clone <repository-url>
cd <repository-folder>
npm install
npm start

On first launch, Joanium walks through setup and stores local app data inside Data/.

Capability Packages

Joanium can now discover integration features from independent workspace packages under Packages/Capabilities/*.

  • Put one product in one package, for example Packages/Capabilities/Notion/.
  • Declare joanium.discovery.features in that package's package.json.
  • Keep the product's connector definition, chat tools, automations, agent data sources, and boot-time wiring inside that package.

That means adding a new product no longer needs a hardcoded core registration path. The product package can own its moving parts and be discovered automatically at startup.

Project Structure

Joanium/
|-- App.js
|-- package.json
|-- Packages/
|   |-- Main/          # Electron-facing services, IPC, paths, windows
|   |-- Renderer/      # SPA shell, pages, shared state, feature modules
|   |-- Automation/    # Scheduler and action execution
|   |-- Agents/        # Scheduled AI jobs and job history
|   |-- Channels/      # External channel responders
|   |-- Connectors/    # AI providers, Gmail, GitHub, and other integrations
|   |-- MCP/           # MCP runtime support
|   `-- System/        # Shared system prompt and app-level logic
|-- Public/            # App shells and static assets
|-- Data/              # Local user data, chats, projects, usage, config
|-- Skills/            # Installed skill definitions
|-- Personas/          # Persona definitions
`-- Docs/              # Architecture and feature documentation

Documentation

The Docs/ folder covers the current runtime and feature set in depth.

Current Status

Joanium is actively being built and sharpened. The foundation is already here: the app shell, project-aware workflows, scheduled runtimes, connector support, and the local-first persistence model that ties everything together.

Built By

Joel Jolly
MIT License

Yorumlar (0)

Sonuc bulunamadi