tensors

A CLI tool for working with safetensor files, CivitAI models, and stable-diffusion.cpp image generation. Supports both local and remote server modes.

Features

https://github.com/user-attachments/assets/2e7629b4-34e7-4cbc-b50e-31d7fdd30239

• Read safetensor metadata - Parse headers, count tensors, extract embedded metadata
• CivitAI integration - Search models, fetch info, identify files by hash
• Download models - Resume support, type-based default paths
• Hash verification - SHA256 computation with progress display
• Image generation - txt2img/img2img via stable-diffusion.cpp server
• Server wrapper - FastAPI wrapper for sd-server with hot reload
• Models database - SQLite cache for local files and CivitAI metadata
• Image gallery - Manage generated images with metadata
• Remote mode - Control remote tsr servers via --remote flag

Installation

# Clone and install
git clone https://github.com/saiden-dev/tensors.git
cd tensors
uv sync

# Or install directly
uv pip install git+https://github.com/saiden-dev/tensors.git

# With server wrapper support
pip install tensors[server]

Usage

Search CivitAI

# Search by query
tsr search "illustrious"

# Filter by type and base model
tsr search -t lora -b sdxl

# Sort by newest, limit results
tsr search -t checkpoint -s newest -n 10

# Filter by tag and period
tsr search --tag anime -p week -b illustrious

# By creator
tsr search -u "username"

# SFW only with commercial use filter
tsr search --sfw --commercial sell

Get Model Info

# Get model info by ID (shows all versions)
tsr get 12345

# Get specific version info
tsr get -v 67890

Download Models

# Download latest version of a model
tsr dl -m 12345

# Download specific version
tsr dl -v 67890

# Download by hash lookup
tsr dl -H ABC123...

# Custom output directory
tsr dl -m 12345 -o ./models

Inspect Local Files

# Read safetensor file and lookup on CivitAI
tsr info model.safetensors

# Skip CivitAI lookup
tsr info model.safetensors --skip-civitai

# Output as JSON
tsr info model.safetensors -j

# Save metadata files
tsr info model.safetensors --save-to ./metadata

Generate Images

Requires a running stable-diffusion.cpp server.

# Generate an image
tsr generate "a cat sitting on a roof"

# Custom size, steps, and output
tsr generate "sunset over mountains" -W 768 -H 512 --steps 30 -o ./output

# Multiple images with seed
tsr generate "cyberpunk city" -b 4 -s 42

# With sampler and negative prompt
tsr generate "portrait" --sampler euler_a -n "blurry, low quality"

Server Wrapper

Manage sd-server process via a REST API. Requires pip install tensors[server].

# Start the wrapper API with a model
tsr serve --model /path/to/model.safetensors

# Custom host and port
tsr serve --model /path/to/model.safetensors --host 0.0.0.0 --port 51200

# Check server status
tsr status

# Hot-reload with a new model
tsr reload --model /path/to/other_model.safetensors

Models Database

Track local safetensor files and cache CivitAI metadata for offline access.

# Scan a directory for safetensor files
tsr db scan /path/to/models

# Link unscanned files to CivitAI by hash
tsr db link

# Cache full CivitAI model data
tsr db cache 12345

# List local files with CivitAI info
tsr db list

# Search cached models offline
tsr db search "pony"
tsr db search -t lora -b sdxl

# Get trigger words for a LoRA
tsr db triggers model.safetensors

# Show database statistics
tsr db stats

Image Gallery

Manage generated images on a remote server.

# List images in gallery
tsr images list --remote junkpile

# Show image metadata
tsr images show IMAGE_ID --remote junkpile

# Download an image
tsr images download IMAGE_ID --remote junkpile -o ./downloads

# Delete an image
tsr images delete IMAGE_ID --remote junkpile

Model Management

List and switch models on a remote server.

# List available models
tsr models list --remote junkpile

# Show active model
tsr models active --remote junkpile

# Switch to a different model
tsr models switch /path/to/model.safetensors --remote junkpile

# List available LoRAs
tsr models loras --remote junkpile

Remote Mode

Control a remote tsr server instead of local operations.

# Configure a remote server
tsr remote add junkpile http://junkpile:51200

# Set default remote
tsr remote default junkpile

# List configured remotes
tsr remote list

# Generate on remote server
tsr generate "a cat" --remote junkpile

# Download model to remote server
tsr dl -m 12345 --remote junkpile

# All commands support --remote flag
tsr status --remote junkpile

Configuration

# Show current config
tsr config

# Set CivitAI API key
tsr config --set-key YOUR_API_KEY

Configuration

Config file: ~/.config/tensors/config.toml

[api]
civitai_key = "your-api-key"

[remotes]
junkpile = "http://junkpile:51200"
local = "http://localhost:51200"

# Optional: set default remote for all commands
default_remote = "junkpile"

Or set API keys via environment variables:

export CIVITAI_API_KEY="your-api-key"      # For CivitAI API access
export TENSORS_API_KEY="your-server-key"   # For server authentication

Default Paths

Data is stored in XDG-compliant paths:

Type	Path
Config	~/.config/tensors/config.toml
Database	~/.local/share/tensors/models.db
Checkpoints	~/.local/share/tensors/models/checkpoints/
LoRAs	~/.local/share/tensors/models/loras/
Gallery	~/.local/share/tensors/gallery/
Metadata	~/.local/share/tensors/metadata/

Search Options

Option	Values
-t, --type	checkpoint, lora, embedding, vae, controlnet, locon
-b, --base	sd14, sd15, sd2, sdxl, pony, flux, illustrious, noobai, auraflow
-s, --sort	downloads, rating, newest
-n, --limit	Number of results (default: 25)
-p, --period	all, year, month, week, day
--tag	Filter by tag (e.g., "anime")
-u, --user	Filter by creator username
--nsfw	none, soft, mature, x
--sfw	Exclude NSFW content
--commercial	none, image, rent, sell
--page	Page number for pagination

Generate Options

Option	Description
-W	Image width (default: 512)
-H	Image height (default: 512)
--steps	Sampling steps (default: 20)
--cfg-scale	CFG scale (default: 7.0)
-s	RNG seed, -1 for random (default: -1)
-b	Batch size / number of images (default: 1)
-n	Negative prompt
-o	Output directory (default: .)
--sampler	Sampler name
--scheduler	Scheduler name
--host	sd-server address (default: 127.0.0.1)
--port	sd-server port (default: 1234)

Server API Endpoints

When running tsr serve, the following endpoints are available:

OpenAPI Documentation: Visit /docs for interactive Scalar API documentation.

Authentication: If TENSORS_API_KEY is set, all endpoints except /status and /docs require authentication via:

• Header: X-API-Key: your-key
• Query param: ?api_key=your-key

Endpoint	Method	Description
/status	GET	Server status (public)
/docs	GET	OpenAPI documentation (public)
/reload	POST	Hot-reload with new model
/api/images	GET	List gallery images
/api/images/{id}	GET	Get image file
/api/images/{id}/meta	GET	Get image metadata
/api/images/{id}/edit	POST	Update image metadata
/api/images/{id}	DELETE	Delete image
/api/models	GET	List available models
/api/models/active	GET	Get active model
/api/models/switch	POST	Switch model
/api/models/loras	GET	List available LoRAs
/api/generate	POST	Generate images
/api/download	POST	Start CivitAI download
/api/db/files	GET	List local files
/api/db/models	GET	Search cached models
/api/db/stats	GET	Database statistics

All sd-server endpoints (/sdapi/v1/*) are proxied through to the underlying process.

Development

# Install dev dependencies
uv sync --group dev

# Run tests
uv run pytest

# Run tests with coverage
uv run pytest --cov=tensors

# Lint and format
uv run ruff check .
uv run ruff format .

# Type check
uv run mypy tensors/

License

MIT