Neighbor Smart Living is an AI-assisted local discovery and merchant operations demo built with Spring Boot and Vue 2. It combines local shop discovery with merchant-facing AI content tools and an optional, token-protected MCP interface.

简体中文版

• Browse local shops and categories, view shop details, and discover nearby shops.
• Explore posts, likes, follows, check-ins, coupons, and flash-sale orders.
• Create and review restaurant reservations while signed in.
• Generate review summaries, reply drafts, blog drafts, coupon copy, and short-video scripts with configurable AI Skills.
• Use the signed-in AI consultant, which can call the application's business tools. Actions such as reservation creation still run through the authenticated user flow.
• Optionally expose read-only shop and voucher lookup tools over MCP. The MCP endpoint is disabled until MCP_API_TOKEN is set.

Browser (Vue 2 + Element UI) -> Nginx -> Spring Boot REST API
                                             |-> MySQL 8
                                             |-> Redis 7 / Redisson
                                             |-> OpenAI-compatible chat API (optional)
                                             `-> MCP HTTP endpoint (optional, bearer token)

Requirements: Docker Desktop with Compose v2. An AI provider key is optional; without one, browsing features still work and AI calls will not return generated content.

cp .env.example .env
# Edit .env and set AI_API_KEY if you want to use chat and AI Skills.
docker compose up --build

Open http://localhost:8080. The API is proxied through the same origin. MySQL is available on 127.0.0.1:3307; Redis is only exposed inside the Compose network.

The first startup initializes the schema and small, synthetic shop/category records. MySQL only runs files under db/mysql when its data volume is empty. To reset this demo database and its Redis data, run docker compose down -v and start it again.

Use Java 17+, Maven 3.9+, MySQL 8+, and Redis 7+. Create a database and apply these files in order:

db/mysql/001-schema.sql
db/mysql/002-demo-data.sql

Set DB_URL, DB_USERNAME, DB_PASSWORD, REDIS_HOST, and REDIS_PORT as needed, then run:

cd AI-dianping-backend
mvn spring-boot:run

For AI features, also set AI_BASE_URL, AI_API_KEY, and AI_MODEL_NAME. MCP_API_TOKEN is optional; without it /mcp responds with HTTP 503. When enabled, use Authorization: Bearer <token>. The MCP tool set contains only read-only shop, category, and voucher lookups.

Run backend unit tests with:

cd AI-dianping-backend
mvn test

Tests are designed to run without a local MySQL or Redis service. The GitHub Actions workflow runs the same Maven test command on Java 17.

Never commit .env or production credentials. Change the development database password before exposing a deployment beyond your machine.

• AI-dianping-backend/: Spring Boot API, business services, AI Skills, consultant tools, and MCP server.
• AI-dianping-frontend/html/hmdp/: static Vue frontend.
• AI-dianping-frontend/conf/nginx.conf: local/Compose Nginx configuration.
• db/mysql/: schema and synthetic demo data; no imported user accounts, phone numbers, or review records.

• SMS login delivery requires an SMS provider configuration; the repository does not include one.
• AI chat and generation require a valid compatible provider key.
• The project has no verified benchmark artifact for the performance percentages sometimes quoted in project descriptions. No such performance claim is made here.
• This is a learning/demo project. Review authentication, rate limiting, storage, and deployment settings before production use.

See CONTRIBUTING.md for setup and pull request guidance, and SECURITY.md for vulnerability reports. See UPSTREAM.md for source attribution. This repository is licensed under the MIT License; see LICENSE. Frontend dependency attributions are listed in THIRD-PARTY-NOTICES.md.