Search with AI

English 中文

Updated: 2024/05/27

Build your own conversation-based search with AI, a simple implementation with Node.js & Vue3. Live Demo

repo: GitHub、 Gitee

Features

• Built-in support for LLM: OpenAI, Google, Lepton, DeepSeek, Ollama(local)
• Built-in support for search engine: Bing, Sogou, Google, SearXNG(Free)
• Customizable pretty UI interface
• Support dark mode
• Support mobile display
• Support local LLM with Ollama
• Support i18n
• Support Continue Q&A with contexts.
• Support Cache results, Force regeneration.
• Support images search. (update date: 2024/05/27)

Deploy with Docker (recommended)

Install Docker.

• Get the code.

git clone https://github.com/yokingma/search_with_ai.git
cd search_with_ai

• Edit .env.docker file.

For a quick start, no need to modify any configuration here, just run docker compose below.

...
# default is for docker-compose, could modify if you need.
OPENAI_KEY=freegpt35
OPENAI_PROXY_URL=http://freegpt35:3040/v1

# Local llm: Ollama hostname, could modify if you need.
OLLAMA_HOST=http://host.docker.internal:11434

# Searxng hostname, could modify if you need.
SEARXNG_HOSTNAME=http://searxng:8080

• run with docker-compose. (No Key required)

this contains SearXNG and FreeGPT3.5 default, just run

docker compose up -d

• or build and run with docker manually. (optional)

docker build -t my_image .
docker run -d -p 3000:3000 --name my_app my_image
# linux run with host network
# docker run -d --network host --name my_app my_image

then visit http://localhost:3000

• Update

Remember to save your .env settings if it has been changed before.

1. run git pull
2. Delete all old containers and images using Docker Desktop or Docker CLI
3. run docker compose down
4. run docker compose up -d

LLM

Support API (needs KEY)

• OpenAI ChatGPT
• Google Gemini
• Lepton LLama2、Mixtral8*7B
• AliYun Qwen
• Baidu Wenxin
• 01.ai
• Moonshot(Kimi)
• DeepSeek
• ChatGLM
• Tencent Hunyuan
• Ollama

Local LLM

Support Ollama, just need run ollama serve.

Search Engine

Built-in support for search engine: Bing, Sogou, Google, SearXNG

SearXNG (Free, No Key required)

install SearXNG with searxng-docker

SearXNG is a free internet metasearch engine which aggregates results from various search services and databases. The service does not track or profile its users, providing online anonymity for those who seek it. Additionally, SearXNG can be used over Tor for online anonymity.

When you install SearxNG, the only active output format by default is the HTML format. You need to activate the json format to use the API. This can be done by adding the following line to the settings.yml file:

search:
    formats:
        - html
        - json

And set limiter to false:

server:
   limiter: false # default is true

You can also set the host in .env:

# SEARXNG_HOSTNAME=<host>

Bing Search

To use the Bing Web Search API, please visit this link to obtain your Bing subscription key.

The Bing Search API is billed, but has a free tier of 1000 calls per month.

Google Search

You have three options for Google Search: you can use the SearchApi Google Search API from SearchApi, Serper Google Search API from Serper, or opt for the Programmable Search Engine provided by Google.

Sogou Search

For users in China.

Step by step setup

Required:

Node.js >= 18

• Server

yarn install && yarn run build

• Web

cd web && yarn install && yarn run build

• Config(.env)

# Bing search key
BING_SEARCH_KEY=
# Google search key
GOOGLE_SEARCH_KEY=
GOOGLE_SEARCH_ID=
# aliyun key
ALIYUN_KEY=
# Yi Key
YI_KEY=
# google gemini
GOOGLE_KEY=
GOOGLE_PROXY_URL=
# baidu
BAIDU_KEY=
BAIDU_SECRET=
# tencent KEY:ID, SECRET:KEY
TENCENT_KEY=
TENCENT_SECRET=
# openai key
OPENAI_KEY=freeduckduckgo
# openai proxy, default is for docker-compose, could modify if you need.
OPENAI_PROXY_URL=http://freeduckduckgo:3456/v1
# deepseek
DEEPSEEK_KEY=#your_key
# chatglm
GLM_KEY=#your_key
# moonshot
MOONSHOT_KEY=
# lepthon key
LEPTON_KEY=
# Local llm: Ollama hostname, could modify if you need.
OLLAMA_HOST=http://host.docker.internal:11434
# Searxng hostname, could modify if you need.
SEARXNG_HOSTNAME=http://searxng:8080
# The count of resources referenced
REFERENCE_COUNT=8
# Whitelist domains, eg. isou.chat,example.org, skip if empty.
WHITELIST_DOMAINS=
# Server Port
PORT=3000
# SearXNG query options, safesearch: Filter search results,  0: None 1: Moderate 2: Strict.
SEARXNG_SAFE=0
# SearXNG query options, language: default is 'all', eg. all/zh/en/en-US/de/it-IT/fr..., this setting has the highest priority.
SEARXNG_LANGUAGE=
# document: https://docs.searxng.org/user/configured_engines.html
SEARXNG_ENGINES=bing,google
SEARXNG_IMAGES_ENGINES=bing
# enable cache, 1 enable, 0 disable
CACHE_ENABLE=1

• Run In the root of the project:

yarn run start 

• Update In the root of the project:

git pull
yarn install
cd web && yarn install && yarn run build

Now you can visit http://localhost:3000