Overview
Seven AI-powered workflows running on n8n, communicating with the AylaniBot Discord bot through webhooks. Each workflow handles a specialized AI task — from context-aware conversations using retrieval-augmented generation to generative art creation and automated content moderation.
Workflows
1. Context-Aware Chat (RAG)
Retrieval-augmented generation using pgvector embeddings for semantic search across the knowledge base. Combines vector similarity with keyword matching to provide contextually relevant responses to user queries.
2. Smart Knowledge Ingestion
Both passive and active modes for building the knowledge base. Passive mode ingests opted-in community messages. Active mode processes uploaded documents, tutorials, and resources into searchable embeddings.
3. AI Summarization
Automated summarization of long threads, documents, and discussions. Configurable output length and focus areas.
4. Generative Art
Image generation via Stable Diffusion through Replicate, with queue management to handle rate limits. Users can generate images through Discord commands with style presets and parameter controls.
5. AI Server Architect
Automated server structure suggestions based on community type and size. Generates channel layouts, role hierarchies, and permission configurations.
6. Custom Role Manager
AI-assisted role creation with gradient color support. Generates role names, colors, and permission sets based on community themes and user activity patterns.
7. AI Moderation
Content moderation powered by Google's Perspective API. Analyzes messages for toxicity, spam, and off-topic content with configurable sensitivity thresholds and automated action triggers.
Key Features
- Seven specialized AI workflows running on self-hosted n8n
- Webhook-based integration with Discord.js bot
- pgvector embeddings for semantic search
- Queue management for rate-limited APIs
- Configurable sensitivity and output parameters per workflow