If you're using macOS, you can install the Replicate CLI using Homebrew:
brew tap replicate/tap
brew install replicate
Or you can build from source and install it with these commands (requires Go 1.20 or later):
make
sudo make install
Grab your API token from replicate.com/account
and set the REPLICATE_API_TOKEN
environment variable.
$ export REPLICATE_API_TOKEN=<your token here>
Usage:
replicate [command]
Core commands:
model Interact with models
prediction Interact with predictions
training Interact with trainings
Alias commands:
run Alias for "prediction create"
train Alias for "training create"
Additional Commands:
completion Generate the autocompletion script for the specified shell
help Help about any command
Flags:
-h, --help help for replicate
-v, --version version for replicate
Use "replicate [command] --help" for more information about a command.
Generate an image with SDXL.
$ replicate run stability-ai/sdxl \
prompt="a studio photo of a rainbow colored corgi"
Prediction created: https://replicate.com/p/jpgp263bdekvxileu2ppsy46v4
Generate an image with SDXL and upscale that image with ESRGAN.
$ replicate run stability-ai/sdxl \
prompt="a studio photo of a rainbow colored corgi" | \
replicate run nightmareai/real-esrgan --web \
image={{.output[0]}}
# opens prediction in browser (https://replicate.com/p/jpgp263bdekvxileu2ppsy46v4)
Fine-tune SDXL with your own images:
$ replicate train --destination mattt/sdxl-dreambooth --web \
replicate/sdxl \
input_images=@path/to/pictures.zip \
use_face_detection_instead=true
# opens the training in browser
Get the schema for SunoAI Bark
$ replicate model schema suno-ai/bark
Inputs:
- prompt: Input prompt (type: string)
- history_prompt: history choice for audio cloning, choose from the list (type: )
- custom_history_prompt: Provide your own .npz file with history choice for audio cloning, this will override the previous history_prompt setting (type: string)
- text_temp: generation temperature (1.0 more diverse, 0.0 more conservative) (type: number)
- waveform_temp: generation temperature (1.0 more diverse, 0.0 more conservative) (type: number)
- output_full: return full generation as a .npz file to be used as a history prompt (type: boolean)
Output:
- type: object