What's New in ReactorAI

Gemini Models Support - Latest Update!

Google Gemini Integration: Access powerful Gemini Pro and Gemini Flash models
Cloud-Based AI: No local setup required - just add your API key
Seamless Switching: Toggle between Ollama (local) and Gemini (cloud) models
Easy Configuration: Simple API key setup in settings
Enhanced Performance: Choose the best model for your specific needs

Update now to experience the power of Google's latest AI models!

Model Context Protocol (MCP)

ReactorAI fully supports the Model Context Protocol (MCP), an open standard that enables AI models to securely interact with local and remote resources. This turns your AI from a text generator into a capable agent that can work with your data and tools.

Capabilities

With MCP, ReactorAI can:

Access Local Files: Read, write, and analyze files on your computer.
Connect to Databases: Query SQL/NoSQL databases directly.
Use External Tools: Execute scripts, call APIs, and automate workflows.
Browse the Web: (Via compatible MCP servers) Search and retrieve live information.

Built-in File System MCP

ReactorAI comes with a pre-installed File System MCP server, ready to use out of the box. No configuration required.

Read & Analyze

Ask ReactorAI to "Analyze the code in my project folder" or "Summarize this PDF". It can traverse directories and read file contents.

Create & Edit

Generate new files, refactor code, or write documentation directly to your disk. "Create a new Python script for..."

Adding Custom MCPs

You can extend ReactorAI with any MCP server, just like you do with Claude Desktop. The configuration is compatible and simple.

Prerequisites: Node.js & uv

Most MCP servers require a runtime environment. Ensure you have these installed:

Node.js: Required for npx servers. Download Here.
uv: Required for Python uv servers. See installation below.

Install uv (macOS/Linux):

curl -LsSf https://astral.sh/uv/install.sh | sh

Install uv (Windows):

powershell -c "irm https://astral.sh/uv/install.ps1 | iex"

Install uv (via pip):

pip install uv

How to Add a Custom MCP:

Click the MCP icon in the chat area to open the menu.
Click the + icon next to MCP Tools.
Enter the configuration JSON (same format as Claude).

Example Configuration (PostgreSQL):

{
  "postgres": {
    "command": "docker",
    "args": [
      "run",
      "-i",
      "--rm",
      "mcp/postgres",
      "postgresql://user:password@localhost:5432/db"
    ]
  }
}

Quick Access from Chat

Manage your MCP servers instantly without leaving your conversation. Click the MCP icon in the chat input area to open the tools menu.

Instant Control

Toggle servers on/off with a single click. Useful for quickly enabling specific tools for a task.

At-a-Glance Status

See which servers are active and how many tools each provides (e.g., "filesystem: 14 tools").

Pro Tip: Shared Ecosystem

Since ReactorAI uses the standard MCP protocol, you can use the vast ecosystem of existing MCP servers created for Claude and other tools!

All Features

Local & Remote Connectivity: Connect to AI models hosted locally (Ollama) or remotely (Ollama/Gemini).
Model Selection: Choose from multiple Ollama models or Gemini models using a dropdown menu.
Session Management: View, rename, search, and manage your chat history.
Session Prompts: Customize system messages to guide the model’s behavior.
Backup & Restore: Export and import chats for safekeeping.
Security: All data stays on your device unless using a remote server.
Multilingual Support: Use the app in multiple languages, including Arabic.

Core Functionality

AI-Powered Chat/Conversation: Interact with AI models through a chat interface.
AI-Powered Image Description/Captioning: Generate descriptions and captions for images using AI.
Code Generation/Display: Generate and display code snippets within the chat.

Effective Prompting Tips

Getting the best results from AI models requires effective prompting. Here are quick tips to improve your interactions:

Quick Tips:

Be Specific: "Write a Python function to sort a list" is better than "help with Python"
Set Context: Explain what you're working on and your experience level
Use Session Prompts: Set consistent behavior with ReactorAI's session prompt feature
Iterate: Refine your prompts based on the responses you get

Want to Master AI Prompting?

Check out our comprehensive prompting guide with advanced techniques, model-specific tips, and real examples.

Read Full Prompting Guide

Screens

Home Screen

Features Overview:

Top Bar:

ReactorAI Logo: Application branding.
Search Bar: Search messages/sessions.
Settings Icon: Access app settings.
Help Icon: Access documentation/help.

Left Sidebar:

Send Dropdown: Select number of last messages for context.
Model Selection: Choose the AI model.
Create New Session: Create and name new chat sessions.
Session History: Manage and restore previous sessions.

Chat Interface:

Image Input/Display: Input and display images.
Request Input: Input prompts.
Send Button: Send messages.
Message Display: Displays user and AI messages.

AI Model Responses: Shows model name.
Copy Button: Copy AI responses.
Delete Button: Delete any message.

Screenshots

Home Screen

Chat Interface

Settings

Model Selector

Using the App

Install and launch the app on your platform
In settings, enter your local or remote Ollama URL, or configure Gemini API access
Select a model and start a new chat session
Use the session prompt input to customize behavior
Manage chats via the sidebar: search, rename, delete
Utilize image analysis by sending images to the chat

Need Help with Ollama Setup?

Check our detailed guide for installing Ollama and downloading models via command line.

View Ollama Installation Guide