AI-powered natural language search and chat for WordPress
FE Search AI adds a conversational AI search experience to WordPress. It indexes your site content, retrieves relevant context with hybrid search, and generates natural language answers grounded in your own posts, pages, and custom post types.
FE Search AI is designed for WordPress sites where visitors need quick, accurate answers from existing content.
- Conversational AI search: Answer visitor questions in natural language.
- RAG architecture: Retrieve relevant content before generating answers.
- Hybrid search: Combine vector search with BM25 keyword ranking for better recall and keyword relevance.
- Multiple AI providers: Use OpenAI, Google Gemini, or Anthropic Claude.
- Embedding support: Generate vectors with OpenAI or Google embedding models.
- Qdrant integration: Store and retrieve vectors with Qdrant Cloud or self-hosted Qdrant.
- Japanese support: Use TinySegmenter or Yahoo! JAPAN Japanese MA API.
- Customizable chat UI: Embed a floating or inline chat interface with a shortcode.
- GitHub Releases updates: Receive updates from GitHub Releases while WordPress.org distribution is pending.
- Developer hooks: Extend providers, prompts, sync behavior, frontend output, and GitHub update behavior.
Version 1.0.0 adds BM25 keyword ranking, GitHub Releases-based updates, separate full sync and real-time sync status tracking, refreshed admin styling, and updated translations.
- Getting Started: Understand the basic setup flow.
- Installation: Install the plugin and confirm requirements.
- Configuration: Configure providers, Qdrant, sync, and chat UI.
- Search Integration: Add the chat/search UI to your site.
- Sync System: Build and maintain the content index.
- Troubleshooting: Resolve common setup and API issues.
- Install and activate FE Search AI.
- Configure API keys for chat, embedding, and optional reranking.
- Connect Qdrant for vector storage.
- Select content types and build the initial index.
- Add
[fe-search-ai]to a page or enable the floating chat UI. - Test visitor questions and adjust prompts, colors, and sync settings.
FE Search AI communicates with external services selected by the site administrator.
| Service | Purpose |
|---|---|
| OpenAI | Chat completions and embeddings |
| Gemini chat completions and embeddings | |
| Anthropic | Claude chat completions |
| Qdrant | Vector storage and retrieval |
| Cohere | Optional reranking |
| Yahoo! JAPAN Japanese MA API | Optional Japanese tokenization |
Review each provider's terms and privacy policy before enabling it.
FE Search AI works as a free plugin, and FE Search AI Pro adds advanced features for production and business use.
- Advanced model selection and model-specific prompts
- Custom OpenAI-compatible endpoints
- Custom field indexing
- Fullscreen chat mode
- User consent and enhanced privacy controls
- External API token management
- MCP server integration for AI agents
- Advanced logging and export features
- Priority support and automatic updates