Overview
A LINE chatbot that brings a pet to life as a personalized AI companion. Built on top of a large language model and the workflow automation tool n8n, it connects through the LINE Messaging API to create an always-available pet account.
The bot supports multi-turn continuous conversation with context memory, image and voice understanding, and can interact via stickers or text. It can also quickly switch to different tones and personalities based on specific business scenarios or needs.
Highlights
- 01
n8n Workflow Automation Architecture
Uses n8n as the core backend integration platform to flexibly orchestrate LINE Webhook events, Gemini API calls, and database storage — eliminating heavy traditional backend development and dramatically improving iteration efficiency.
- 02
Multi-turn Context Memory (Context Window)
Distinguishes sessions and permissions by user ID, enabling precise and independent personalized conversation contexts.
- 03
Image & Voice Understanding (Multimodal Agent)
The bot can understand image and audio content and interact accordingly.
- 04
Structured Prompt Engineering
Carefully tuned system prompts to ensure conversations align with a specific character persona and conversational style.
- 05
Long-term Memory (AI Knowledge Base)
The pet bot retains memories of all past events. Past memories are stored as JSONL, vectorized (embedded) into a semantic search database, and exposed to the AI via tool calls — making the pet bot feel genuinely lifelike.
- 06
Lightweight Cloud Deployment (Docker & VPS)
n8n and related services are containerized with Docker and deployed on a DigitalOcean VPS.