A chatbot that reads the graph before it speaks

What it does

This is a browser-based chatbot frontend backed by a knowledge-graph API. Clone it, open index.html, and you’re chatting with a bot that answers “姚明多高啊？” by consulting structured knowledge rather than pattern matching alone. The project also exposes a simple GET/POST API for building your own integrations.

The interesting bit

The author isn’t hiding the plumbing — the bot is explicitly the consumer end of a pipeline that includes a separate NLP tool (Jiagu), knowledge graph (KnowledgeGraph), and semantic understanding module (Semantic). It’s a rare case of a chatbot repo that admits it’s the tip of a spear, not the whole weapon.

Key highlights

• Zero-build setup: static HTML/CSS/JS, runs locally without a server
• Free tier API with appid registration at console.ownthink.com
• Supports both GET and POST to api.ownthink.com/bot
• Companion projects cover the full stack from tokenization to triple stores
• Live demo at ownthink.com/robot.html

Caveats

• The README is entirely in Chinese; English speakers will need translation help
• API docs are minimal — error handling, rate limits, and data retention are unspecified
• The “CSS” language tag is misleading; this is a thin frontend with the smarts living server-side

Verdict

Worth a look if you’re building Chinese-language conversational AI and want a working reference architecture to dissect. Skip it if you need self-hosted intelligence or comprehensive English documentation.