← Plugin Directory / AI Chat Plugin
🔌
FREE

AI Chat Plugin

The AI Chat Plugin is a comprehensive, multimodal conversational assistant built specifically for BinaryCMS. It transforms the CMS into an interactive workspace.

About this Plugin

The AI Chat Plugin is a comprehensive, multimodal conversational assistant built specifically for BinaryCMS. It transforms the CMS into an interactive workspace by providing a persistent, AI-powered chat interface in both the administrative backend (“AI Chat Studio”) and the public-facing frontend via a floating widget.

Powered by OpenRouter API integration, it allows users to dynamically switch between cutting-edge LLMs (like GPT-4o, Claude 3.5 Sonnet, and Llama 3) on the fly, while supporting advanced multimodal capabilities including text, file parsing, image vision, and voice inputs.

Features

  1. Multimodal Interactions Image Vision: Attach images to your prompts to have the AI analyze, describe, or extract text from visual data.

Voice Notes (Speech-to-Text): Use the built-in microphone integration to record voice notes directly in the browser and send them to native audio-processing models (e.g., gpt-4o-audio-preview).

File Parsing: Upload text-based files (.txt, .py, .json, .csv, etc.). The plugin extracts the raw text client-side and injects it securely into the prompt, ensuring 100% compatibility across all AI models, even those without native file support.

  1. AI Chat Studio (Backend) Persistent Conversation History: All chats are automatically saved to the SQLite database. Access your past conversations anytime from the sidebar.

Inline Chat Management: Hover over any chat in the sidebar to reveal a quick-delete trash icon, or use the Options dropdown to rename chats for better organization.

Model Overrides: Easily override the default global AI model on a per-conversation basis. Use Claude for coding in one tab, and ChatGPT for copywriting in another.

Syntax Highlighting & Markdown Rendering: Code blocks are automatically formatted with Prism.js syntax highlighting. Each code block features a convenient one-click “Copy” button.

  1. Public Frontend Widget Live Support Interface: Enable the frontend widget to provide public users with an AI assistant on your live website.

Customizable Aesthetics: Tailor the widget’s primary color and default welcome message through the backend settings to match your brand identity.

Shortcode Integration: Can be conditionally rendered or managed via CMS shortcodes or layout injection.

  1. Admin Settings & Configuration Global System Prompts: Define the core personality and rules for your AI assistant across all new chats.

Temperature & Token Limits: Fine-tune the AI’s creativity and response length.

Rate Limiting: Protect your OpenRouter API credits by setting maximum requests per minute.

Context Window Management: Configure the maximum number of historical messages sent with each new prompt to optimize token usage.

Technical Architecture Backend (Go / SQLite) Data Models: Stores conversations and messages in robust SQLite tables (ai_conversations, ai_messages).

Payload Handling: Converts complex multi-part messages (Base64 images, audio blobs) into standardized OpenRouter JSON payloads (ContentPart struct) dynamically.

Secure Isolation: Runs entirely as an isolated plugin inside BinaryCMS without modifying core framework files.

Frontend (Vanilla JS / Bootstrap)

Responsive Design: Built using Bootstrap 5 utility classes for a clean, modern, and mobile-friendly interface.

Client-Side Media Processing: Utilizes the native FileReader API for images/documents and MediaRecorder API for WebRTC microphone capture, offloading processing from the server.

Optimistic UI: Messages appear instantly in the chat window while the background request processes, ensuring a snappy user experience.

Setup & Usage

API Key: Ensure you have configured your OpenRouter API key in the AI Provider Settings plugin (this chat plugin hooks into the core AI provider).

Settings: Navigate to AI Chat > Settings to configure your System Prompt, Default Model, and Frontend Widget preferences. Frontend Display: Check “Enable Frontend Widget” in the settings. The chat bubble will automatically appear in the bottom-right corner of your live website.

Interacting: In the AI Chat Studio, click the Paperclip icon to attach files or the Microphone icon to record audio. Hit Ctrl Enter or click Send to communicate with the AI.