Skip to main content
The Remyx CLI is the command-line companion to Remyx: list and manage models, run MyxBoard evaluations, deploy with Docker, work with datasets, and search the research catalog. Source and the latest command definitions live in the official repo.

​
Requirements

  • Python 3 and pip
  • A Remyx API key (get it here)
  • For deploy: Docker, Docker Compose, and (for GPU) the NVIDIA Container Toolkit, as described in the repo

​
Installation

pip install remyxai
Verify the executable is on your PATH: 
remyxai --help

​
Authentication

Remyx authenticates API calls with a bearer token. Export it before running commands that hit the API: 
export REMYXAI_API_KEY=<your-key-here>
The client uses this variable when calling https://engine.remyx.ai/api/v1.0. Do not commit keys; use your shell profile, .env (with a loader), or your CI secret store.

​
Command overview

AreaCLI entry points
Modelslist_models, summarize_model
Evaluationsevaluate_myxboard
Deploydeploy_model
Datasetsdataset
Searchsearch (subcommands below)
Run remyxai <command> --help for options. Full detail: CLI repo.

​
Models

​
List models

remyxai list_models

​
Summarize a model

remyxai summarize_model <model-name>

​
Delete or download a model (Python)

The CLI may not expose every model operation; use the API modules when needed: 
from remyxai.api.models import delete_model, download_model

# delete_model("<model-name>")
# download_model("<model-name>", "onnx")  # or other supported format

​
MyxBoard evaluations

Compare models on tasks such as myxmatch or benchmark (see EvaluationTask in the repo). 
# Models first, then task names (see repo for valid task tokens)
remyxai evaluate_myxboard org/model-a org/model-b myxmatch benchmark

​
Deploy (local Docker)

Brings up or tears down a deployment using Docker Compose (after downloading the deployment package from Remyx). 
remyxai deploy_model "<model-name>" up
remyxai deploy_model "<model-name>" down

​
Datasets

​
List (CLI)

remyxai dataset list

​
Delete or download (Python)

The HTTP API expects a dataset type and dataset name: 
from remyxai.api.datasets import delete_dataset, download_dataset

# delete_dataset("<dataset_type>", "<dataset_name>")
# download_dataset("<dataset_type>", "<dataset_name>")
If your CLI build wires dataset delete / dataset download to the same signatures, use those; otherwise use the Python calls above. Check the repo for the latest dataset CLI behavior.

​
Search (papers & Docker assets)

remyxai search query "natural language query" --max-results 10
remyxai search info <arxiv-id> --format text
remyxai search list --limit 20 --offset 0
remyxai search stats
Optional thin wrapper: from remyxai.client.search import SearchClient.

​
Inference (local Triton)

After a model is served (for example on localhost:8000), run inference via remyxai.api.inference.run_inference or RemyxAPI.run_inference — see the repo for parameters and tensor names.

​
Troubleshooting

  • REMYXAI_API_KEY missing or invalid — Export the key in the same shell session you use for remyxai.
  • Deploy fails — Confirm Docker/Compose and GPU runtime requirements from the repo.
  • Import errors in Python — Prefer explicit imports (remyxai.api.models, remyxai.api.tasks, …) if from remyxai.api import … is not available in your version.

​
Next steps

  • Browse commands and examples in the Remyx CLI repository.
  • Keep your API key scoped to the environment where you develop and deploy.