Magic: Open Source AI Productivity Platform to Help Enterprises Efficiently Build Intelligent Applications

Magic is the world's first open source one-stop AI productivity platform designed to help enterprises and individuals work efficiently through AI technology. It integrates several modules, including a universal AI agent, AI workflow orchestration, and an enterprise instant messaging tool.Magic adopts an open-source philosophy, with its code publicly available on GitHub, and can be accessed and customized by anyone for free. The platform is suitable for working professionals, entrepreneurs and enterprise managers to enhance work efficiency through intelligent tools and help enterprises rapidly deploy AI applications.Magic supports multiple large-scale language models, is easy to operate and powerful, and is suitable for a variety of complex task scenarios.

Function List

• Super Magic (Super Magic): A general-purpose AI agent that supports complex task processing, such as automated data analysis, content generation, etc.
• Magic Flow: Visual AI workflow orchestration tool that allows users to design and execute automated workflows.
• Magic IM: Enterprise-grade instant messaging system with integrated AI agents to support team collaboration and intelligent conversations.
• Teamshare OS (coming soon): An enterprise-class online collaborative office system that provides document management and team collaboration features.
• Support for multiple large-scale language models: Compatible with models in OpenAI format, allowing users to flexibly configure AI capabilities.
• Open Source Support: Full code is available via GitHub to support developer contributions and custom development.

Using Help

Installation process

Magic has a simple installation process for users with a basic technical background. Here are the detailed installation steps, based on the official GitHub documentation:

1. Cloning Codebase
   Run the following command in the terminal to clone the Magic code locally:

   git clone https://github.com/dtyq/magic.git
   cd magic
   

   This will download Magic's full codebase to your local device.

2. Configuring Environment Variables
   Magic needs to be configured with at least one large language model that supports the OpenAI format (e.g. ChatGPT or similar). Copy the example configuration file:

   cp .env.example .env
   cp config/.env_super_magic.example .env_super_magic
   

   show (a ticket).envcap (a poem).env_super_magicfile, fill in your language model API key and other necessary parameters. Make sure to configure at least one language model, otherwise Magic won't work properly.

3. Starting services
   • Front-end activation: Run./bin/magic.sh start, the service will run in the foreground and is suitable for debugging.
   • backstage activation: Run./bin/magic.sh daemonThe service runs in the background and is suitable for production environments.
   • status: Run./bin/magic.sh status, to see if the service is running properly.
   • View Log: Run./bin/magic.sh logs, check the run log to troubleshoot the problem.
4. Accessing the RabbitMQ Management Interface
   Magic uses RabbitMQ to manage task queues. Accesshttp://localhost:15672If you want to use the default username, you have to use the default username.adminand passwordsmagic123456Log in and check the status of your tasks.
5. Enterprise Functionality Configuration
   If you are an enterprise user, you can contact the Magic team (email: dev@dtyq.com) for support on additional administrative features such as advanced privilege control or customized deployment.

Function Operation Guide

Super Magic (Super Magic)

SuperMagic is the core AI agent of the Magic platform, suitable for handling complex tasks. Users can interact with the AI via text input or API calls. Example:

• Task automationEnter "Analyze last month's sales data and generate a report" and the AI will automatically process the data and output a structured report.
• Content generationEnter "Write 100 words of promotional copy for a new product" and the AI will quickly generate copy that meets the requirements.
• workflow::
  1. Log in on the Magic platform and go to the SuperMagic module.
  2. Describe the task in the input box, e.g., "Organize data in an Excel spreadsheet and extract key metrics".
  3. Set the output format (text, table or JSON).
  4. Click "Execute" and the AI will return the result in a few seconds.
  5. Batch task processing can be realized through API calls, suitable for enterprise-level applications.

Magic Flow

Magic Flow is a visual workflow orchestration tool for users who need to automate repetitive tasks. Users can design workflows through a drag-and-drop interface. Example:

• Creating Workflows::
  1. Log in to the Magic platform and go to the Magic Flow module.
  2. Click "New Workflow" and select the trigger condition (e.g. "Receive new email").
  3. Add action nodes such as "Extract email content" or "Store data in database".
  4. Connect the nodes to form a complete process.
  5. Save and run, and the system will automatically execute the workflow.
• sample scenario: Set up a workflow to automatically pull order information and update the CRM system when a customer email is received.

Magic IM

Magic IM is an enterprise-grade instant messenger with integrated AI agents. Team members can use it to communicate in real-time and invoke AI to accomplish tasks. Example:

• intelligent dialog: Type "@Magic summarize meeting minutes" in the chat window and the AI will automatically generate a summary of the meeting.
• Teamwork: Support for creating groups, sharing files and assigning tasks.
• workflow::
  1. Log in to the Magic platform and go to the Magic IM module.
  2. Create team groups and invite members to join.
  3. Call AI using @Magic in the dialog box and enter the task command.
  4. View the results returned by the AI or share them directly with team members.

open source contribution

Magic encourages developers to participate in the open source community. Users can:

• interviewshttps://github.com/dtyq/magic, check out the contribution guidelines.
• Submit code, report problems, or suggest new features.
• If you find a security vulnerability, you can send an email todev@dtyq.com, the team will deal with it in a timely manner.

caveat

• Environmental requirements: Magic needs to run on a Linux or Unix-like system, Ubuntu 20.04 or higher is recommended.
• network requirement: Ensure that the device is connected to a stable network, some functions require access to external APIs.
• documentation support: Detailed documentation can be accessed athttps://docs.letsmagic.cn/en, contains API descriptions and advanced configuration guides.

application scenario

1. Automation for small and medium-sized enterprises
   Small and medium-sized businesses use Magic Flow to automate routine tasks such as processing customer orders, generating sales reports, and saving on labor costs.
2. Teamwork and Communication
   Remote teams communicate in real-time via Magic IM, combining with AI agents to quickly complete document organization or meeting minutes to improve collaboration efficiency.
3. Developers customize AI apps
   Developers use Magic's open source code to customize exclusive AI applications, such as developing intelligent customer service systems for e-commerce platforms.
4. Data analysis and reporting
   Data analysts use SuperMagic to work with complex datasets, automatically generate visual charts, and reduce analysis time.

QA

1. Is Magic free?
   Magic is an open source platform and is free to use for individuals and developers. Enterprise users can contact the team for paid premium feature support.
2. How do I configure a large language model?
   exist.envcap (a poem).env_super_magicFill in the file with the model API key that supports the OpenAI format, save it and restart the service.
3. What operating systems does Magic support?
   At present, the main support for Linux and Unix-like systems, recommended Ubuntu 20.04 or above, does not support Windows direct deployment.
4. How can I participate in open source contributions?
   interviewshttps://github.com/dtyq/magic, read the contribution guidelines, submit code or issue feedback.