🇬🇧 English ⇌ 🇨🇳 中文

ImgSearch is a lightweight reverse image search engine that supports search by image or text. Built on TinyCLIP and HNSWlib, it's fast and resource-efficient, running on devices with just 2GB of RAM. Use it standalone or integrate it as a Python library.

• Search by image: Upload a query image to find similar ones quickly
• Search by text: Find images matching natural language descriptions
• Image similarity comparison: Compute similarity scores (0-100%) between two images
• Batch image addition: Add single files or folders (recursive), skipping duplicates automatically
• Multi-database support: Create and manage multiple independent image libraries
• Similarity threshold filtering: Filter search results by minimum similarity (e.g., ≥80%)

To use the full functionality of ImgSearch (including client and server), you need to specify the all dependency group during installation. This will install all the dependency packages required to run the server, such as PyTorch, on the system.

pip install 'imgsearch[all]'

The standard installation includes only the client-side dependencies:

pip install imgsearch

⚠️ Note: Since the standard installation lacks PyTorch and other dependencies, you cannot run the isearch service start command.

For users in non-CUDA environments who want to run the server, it is recommended to install the CPU version of PyTorch first, and then install ImgSearch:

# Install CPU version of PyTorch
pip install torch torchvision --index-url https://download.pytorch.org/whl/cpu

# Install ImgSearch
pip install 'imgsearch[all]'

If using uv as your Python package manager:

uv pip install --torch-backend cpu imgsearch

ImgSearch follows a client-server architecture: the server handles indexing and search tasks, while the client manages user requests. Before using ImgSearch, start the server process.

The server supports unix domain sockets (default for local, efficient connections) or TCP binding. The default unix socket is at ~/.isearch/isearch.sock.

The default model is ViT-45LY (TinyCLIP-auto-ViT-45M-32-Text-18M-LAIONYFCC400M). See Model Selection Guide for available options.

# Start ImgSearch server with default settings
isearch service start

• -B BIND: Server binding, either UDS path or IP:PORT (e.g., -B /path/to/isearch.sock or -B 127.0.0.1:5000)
• -b BASE_DIR: Index database directory (default: ~/.isearch/)
• -m MODEL_KEY: Model name (see Model Selection Guide)
• -L LOG_LEVEL: Log level (debug, info, warning, error, critical; default: info)

isearch service stop

isearch service status

Example output:

iSearch service is running
 - PID  : 86269
 - MEM  : 1301.6 MB
 - Base : /home/seamile/.isearch
 - Model: ViT-45LY

ImgSearch can run as a background system service that starts and stops with your system. On Linux, it uses systemd; on macOS, launchd. The tool auto-detects your environment — no manual configuration needed.

The setup command creates and starts the ImgSearch system service. It accepts the same optional parameters as the start command.

⚠️ Note: This downloads model files to the cache directory ~/.cache/clip. Most models range from 100 MB to 230 MB—ensure you have enough disk space.

isearch service setup

The remove command stops and uninstalls the ImgSearch system service. It does not delete database files.

isearch service remove

The add command indexes images in the specified database. It supports formats like jpg, jpeg, png, bmp, and webp. Folders are scanned recursively, with duplicates automatically filtered (based on labels).

# Add single files or folders to the default database
isearch add ./images/photo1.jpg ./pictures/

# Use filename as label (default: absolute path)
isearch add -l name ./images/

# Specify database and binding
isearch add -d my_gallery ./photos/ -B ./isearch.sock

The add command extracts 512-dimensional feature vectors from images using TinyCLIP and stores them in an HNSW index.

The search subcommand handles image searches, but for ease of use, it's the default—omit it if your arguments don't match other commands.

Search command: isearch [search] QUERY.

# Search similar images, return top 10 results (similarity ≥0%)
isearch ./query.jpg

# Equivalent:
isearch search ./query.jpg

# Set minimum similarity threshold and result count
isearch -n 5 -t 80 ./query.jpg

# Automatically open result images (the Label when adding an image must be a path)
isearch -o ./query.jpg

# Search for "red flower" related images
isearch "red flower"

# Specify count and threshold
isearch -n 3 -t 70 "sports car"

Results sorted by similarity descending, show path and percentage. Example:

Searching sports car...
Found 5 similar images (similarity ≥ 70.0%):
 1. /path/to/img1.jpg	92.3%
 2. /path/to/img2.png	85.1%
 3. /path/to/img3.jpg	78.4%

isearch db -l

Output:

Available databases:
* default (loaded)
* gallery
* test_db

isearch db -i my_db

Example output:

Database "my_db"
 - Base: /home/seamile/.isearch
 - Capacity: 150000
 - Count: 141610
 - Size: 313.9 MB

isearch db my_db --delete /path/to/img1.jpg /path/to/img2.jpg ...

# Confirm then clear
isearch db --clear my_db

# Confirm then drop
isearch db --drop my_db

⚠️ Warning: All delete operations are irreversible. --clear deletes all data in the database, --drop deletes the database file.

isearch cmp ./img1.jpg ./img2.png

Output:

Similarity between images: 87.5%

ImgSearch can be used as a Python module in other projects.

from imgsearch.client import Client

# Create client (connects to local service by default)
cli = Client(db_name='default', bind='~/.isearch/isearch.sock')

# Add images (returns count)
image_paths = ['./img1.jpg', './img2.png', './folder/']
n_added = cli.add_images(image_paths, label_type='path')  # or 'name'
print(f'Added {n_added} images for processing')

# Search by image (returns [(path, similarity%), ...] or None)
results = cli.search('./query.jpg', num=5, similarity=80)
if results:
    for path, sim in results:
        print(f"{path} (similarity: {sim}%)")
else:
    print('No matching results or search queue full')

# Search by text
results = cli.search('red apple', num=10, similarity=0)
for path, sim in results:
    print(f"{path} (similarity: {sim}%)")

# Compare similarity (returns 0-100 float)
similarity = cli.compare_images('./img1.jpg', './img2.jpg')
print(f'Similarity: {similarity}%')

# Database operations
dbs = cli.list_dbs()
print(f'Available databases: {dbs}')

info = cli.get_db_info()
print(f'Database information: {info}')

# Clear database (returns True/False)
cleared = cli.clear_db()
print(f'Clear success: {cleared}')

⚠️ Note: Start the service first (isearch service start) for module usage, or connections will fail.

ImgSearch supports various TinyCLIP models, with the default ViT-45LY offering a good balance of speed, accuracy, and resource use for most cases.

Data source: TinyCLIP Model Zoo. Use smaller models like ViT-8Y on low-end devices; opt for larger ones like ViT-63LY for precision-focused tasks.

To switch models, restart the service:

# Restart ImgSearch service
isearch service stop
isearch service start -m NEW_MODEL_KEY

🇨🇳 中文 ⇌ 🇬🇧 English

ImgSearch 是一款轻量级的反向图片搜索引擎，支持以图搜图和文字描述搜图。基于 TinyCLIP 和 HNSWlib 构建，速度快、资源占用低，可在 2GB 内存设备上运行。可作为独立搜索引擎使用，或作为 Python 库集成到其他系统。

• 以图搜图：上传查询图片，快速找到相似图像
• 文字搜图：通过自然语言描述搜索相关图片
• 图像相似度比较：计算两张图片的相似度分数（0-100%）
• 批量添加图片：支持单个文件或文件夹（递归添加），自动跳过重复
• 多数据库支持：可创建和管理多个独立图片库
• 相似度阈值过滤：搜索结果可设置最小相似度（如 ≥80%）

要使用 ImgSearch 的完整功能（包含客户端和服务端），需要在安装时指定 all 依赖组，这会将服务端运行所需的 PyTorch 等依赖包全部安装到系统上：

pip install 'imgsearch[all]'

标准安装仅包含客户端所需的依赖：

pip install imgsearch

⚠️ 注意：由于标准安装缺少 Pytorch 等依赖项，所以无法执行 isearch service start 命令。

对于非 CUDA 环境的用户如果要运行服务端时，建议先安装 CPU 版 PyTorch，然后再安装 ImgSearch：

# 安装 CPU 版 PyTorch
pip install torch torchvision --index-url https://download.pytorch.org/whl/cpu

# 安装 ImgSearch
pip install 'imgsearch[all]'

使用 uv 作为 Python 包管理工具的可以这样安装：

uv pip install --torch-backend cpu 'imgsearch[all]'

ImgSearch 整体为 C-S 架构，服务端处理索引和搜索，客户端处理用户请求。使用前需先开启他的服务端程序。

ImgSearch 服务端支持 Unix 域套接字或 TCP 绑定。默认使用 Unix 域套接字运行在本地，稳定且高效。默认 UDS 位置为 ~/.isearch/isearch.sock。

默认模型为 ViT-45LY（TinyCLIP-auto-ViT-45M-32-Text-18M-LAIONYFCC400M），可用模型列表详见 模型选择指南。

# 使用默认参数启动 Img Search 服务
isearch service start

• -B BIND: 服务端绑定方式，可选 UDS 或 IP:PORT 两种格式。如：-B /path/to/isearch.sock 或 -B 127.0.0.1:5000
• -b BASE_DIR: 索引数据库目录，默认为 ~/.isearch/
• -m MODEL_KEY: 模型名称，可选项详见 模型选择指南
• -L LOG_LEVEL: 日志级别，可选 debug、info、warning、error、critical。

isearch service stop

isearch service status

# 输出：
iSearch service is running
 - PID  : 86269
 - MEM  : 1301.6 MB
 - Base : /home/seamile/.isearch
 - Model: ViT-45LY

ImgSearch 可以作为系统服务在后台运行，并可以随系统自动启动和停止。在 Linux 上，ImgSearch 使用 systemd 来管理守护进程，在 macOS 上则使用 launchd。ImgSearch 会自动识别运行环境，无需手动指定服务管理程序。

setup 命令用于创建和启动 ImgSearch 的系统服务。它的可选参数与 start 命令 一样。

⚠️ 注意：此操作会下载模型文件到缓存目录 ~/.cache/clip，大部分模型的大小在 100 MB 至 230 MB 之间，请确保磁盘空间充足。

isearch service setup

remove 命令用于停止和卸载 ImgSearch 的系统服务，但 不会删除 数据库文件。

isearch service remove

add 命令用于将图片添加到指定数据库。支持 jpg, jpeg, png, bmp, webp 等格式。文件夹会递归添加所有图片，自动过滤重复项（基于 Label）。

# 添加单个文件或文件夹到默认数据库
isearch add ./images/photo1.jpg ./pictures/

# 使用文件名作为标签（默认：绝对路径）
isearch add -l name ./images/

# 指定数据库和绑定
isearch add -d my_gallery ./photos/ -B ./isearch.sock

add 命令会通过 TinyCLIP 提取图片的特征（512 维向量），并存储在 HNSWlib 索引中。

搜索图片使用 search 子命令，为了操作方便，isearch 已将它设为默认子命令，使用时可以省略。

搜图方式：isearch [search] QUERY。

# 搜索相似图片，默认返回前 10 个结果
isearch ./query.jpg

# 等价于：
isearch search ./query.jpg

# 设置最小相似度阈值和结果数量
isearch -n 5 -t 80 ./query.jpg

# 自动打开结果图片（添加图片时的 Label 必须是 path）
isearch -o ./query.jpg

# 搜索 "red flower" 相关图片
isearch "red flower"

# 指定数量和阈值
isearch -n 3 -t 70 "sports car"

搜索结果按相似度降序排列，显示路径和相似度。示例：

Searching sports car...
Found 3 similar images (similarity ≥ 70%):
 1. /path/to/img1.jpg	92.3%
 2. /path/to/img2.png	85.1%
 3. /path/to/img3.jpg	78.4%

isearch db -l

输出：

Available databases:
* default (loaded)
* gallery
* test_db

isearch db -i my_db

输出示例：

Database "my_db"
 - Base: /home/seamile/.isearch
 - Capacity: 150000
 - Count: 141610
 - Size: 313.9 MB

isearch db my_db --delete /path/to/img1.jpg /path/to/img2.jpg ...

# 确认后清空
isearch db --clear my_db

# 确认后删除
isearch db --drop my_db

⚠️ 警告：所有删除操作均不可逆，--clear 会删除数据库中所有数据，--drop 会删除数据库文件。

isearch cmp ./img1.jpg ./img2.png

输出：

Similarity between images: 87.5%

ImgSearch 可作为 Python 模块导入到其他项目中。

from imgsearch.client import Client

# 创建客户端（默认连接本地服务）
cli = Client(db_name='default', bind='~/.isearch/isearch.sock')

# 添加图片（返回添加数量）
image_paths = ['./img1.jpg', './img2.png', './folder/']
n_added = cli.add_images(image_paths, label_type='path')  # 或 'name'
print(f'已添加 {n_added} 张图片进行处理')

# 以图搜图（返回 [(路径, 相似度%), ...] 或 None）
results = cli.search('./query.jpg', num=5, similarity=80)
if results:
    for path, sim in results:
        print(f"{path} (相似度: {sim}%)")
else:
    print('无匹配结果或搜索队列满')

# 以文字搜图
results = cli.search('red apple', num=10, similarity=0)
for path, sim in results:
    print(f"{path} (相似度: {sim}%)")

# 比较相似度（返回 0-100 float）
similarity = cli.compare_images('./img1.jpg', './img2.jpg')
print(f'相似度: {similarity}%')

# 数据库操作
dbs = cli.list_dbs()
print(f'可用数据库: {dbs}')

info = cli.get_db_info()
print(f'数据库信息: {info}')

# 清空数据库（返回 True/False）
cleared = cli.clear_db()
print(f'清空成功: {cleared}')

⚠️ 注意：模块使用需先启动服务（isearch service start），否则连接失败。

ImgSearch 支持多种 TinyCLIP 模型，默认的 ViT-45LY 平衡了速度、准确率和资源占用，适用于大多数场景。

数据来源：TinyCLIP 模型库。选择小模型（如 ViT-8Y）用于低端设备；大模型（如 ViT-63LY）用于高精度需求。

切换模型需重启服务：

# 重启 isearch 服务
isearch service stop
isearch service start -m NEW_MODEL_KEY