clarifai-cli

name	clarifai-cli
description	Execute Clarifai operations using the `clarifai` CLI. Use when the user needs to authenticate, check identity, manage apps, compute clusters, nodepools, deployments, pipelines, pipeline runs, artifacts, or initialize models. Covers model init (toolkits, MCP, GitHub templates), pipeline templates, compute orchestration CLI, and all model/pipeline operations.

Clarifai CLI

IMPORTANT: For SDK methods and CLI commands and code generation, first scan references/sdk-cli-reference.md

The clarifai CLI provides terminal access to all Clarifai platform operations including model initialization, upload, local testing, pipeline management, and compute orchestration.

Quick Command Reference

Task	Command
Login	clarifai login
Logout	clarifai logout
Whoami	clarifai whoami
Config profiles	clarifai config
Initialize model	clarifai model init
Init from toolkit	clarifai model init --toolkit vllm
Init MCP server	clarifai model init --toolkit mcp
Upload model	clarifai model upload ./model
Serve locally	clarifai model serve ./model
Serve as gRPC	clarifai model serve ./model --grpc --port 8000
Deploy to cloud	clarifai model deploy ./model
Deploy with GPU	clarifai model deploy ./model --instance gpu-nvidia-a10g
Deploy existing model	clarifai model deploy --model-url <url>
Check status (by deployment)	clarifai model status --deployment <id>
Check status (by model)	clarifai model status --model-url <url>
Stream logs	clarifai model logs --deployment <id>
Undeploy	clarifai model undeploy --deployment <id>
Predict	clarifai model predict <model_ref> "Hello!"
Predict with inputs	clarifai model predict --model-url <url> -i prompt="Hi" -i max_tokens=100
List models	clarifai model list
List GPU instances	clarifai list-instances
Filter instances	clarifai list-instances --cloud aws --gpu A10G
Init pipeline	clarifai pipeline init
Init from template	clarifai pipeline init --template classifier-pipeline-resnet
Upload pipeline	clarifai pipeline upload
Run pipeline	clarifai pipeline run --compute_cluster_id <id> --nodepool_id <id>
List pipelines	clarifai pipeline list
List pipeline templates	clarifai pipelinetemplate list
Discover templates	clarifai pipelinetemplate discover
Manage pipeline runs	clarifai pipelinerun
List deployments	clarifai deployment list --compute_cluster_id <id>
Create deployment	clarifai deployment create <deployment_id> --compute_cluster_id <id> --nodepool_id <id>
Delete deployment	clarifai deployment delete <deployment_id>
List nodepools	clarifai nodepool list --compute_cluster_id <id>
Create nodepool	clarifai nodepool create <nodepool_id> --compute_cluster_id <id>
Delete nodepool	clarifai nodepool delete <nodepool_id>
List apps	clarifai app list [--user_id <user_id>]
Create app	clarifai app create <app_id>
Delete app	clarifai app delete <app_id>
List compute clusters	clarifai computecluster list
Create compute cluster	clarifai computecluster create <cluster_id> --config <config_file>
Delete compute cluster	clarifai computecluster delete <cluster_id>
Manage artifacts	clarifai artifact
Run script with context	clarifai run <script.py>
Shell completion	clarifai shell-completion

Use Case Routing

User Request	Primary Command	Follow-up
"Check who I'm logged in as"	clarifai whoami
"Log out"	clarifai logout
"Manage config profiles"	clarifai config
"List my apps"	clarifai app list
"List apps for another user"	clarifai app list --user_id <user_id>
"Create a new app"	clarifai app create <app_id>
"Delete an app"	clarifai app delete <app_id>
"List compute clusters"	clarifai computecluster list
"Create compute cluster"	clarifai computecluster create <id> --config <file>
"Delete compute cluster"	clarifai computecluster delete <id>
"List nodepools"	clarifai nodepool list --compute_cluster_id <id>
"Create nodepool"	clarifai nodepool create <id> --compute_cluster_id <id>
"Delete nodepool"	clarifai nodepool delete <id>
"List deployments"	clarifai deployment list --compute_cluster_id <id>
"Create deployment"	clarifai deployment create <id> --compute_cluster_id <id> --nodepool_id <id>
"Delete deployment"	clarifai deployment delete <id>
"List pipeline runs"	clarifai pipelinerun list
"Pause pipeline run"	clarifai pipelinerun pause <run_id>
"Resume pipeline run"	clarifai pipelinerun resume <run_id>
"Cancel pipeline run"	clarifai pipelinerun cancel <run_id>
"Discover pipeline templates"	clarifai pipelinetemplate discover
"List pipeline templates"	clarifai pipelinetemplate list
"Manage artifacts"	clarifai artifact
"Deploy vLLM model"	clarifai model init --toolkit vllm	clarifai model deploy
"Deploy SGLang model"	clarifai model init --toolkit sglang	clarifai model deploy
"Deploy HuggingFace model"	clarifai model init --toolkit huggingface	clarifai model deploy
"Deploy Ollama model"	clarifai model init --toolkit ollama	clarifai model serve
"Create MCP server"	clarifai model init --toolkit mcp	clarifai model deploy
"Create OpenAI-compatible model"	clarifai model init --toolkit openai	clarifai model deploy
"Train image classifier"	clarifai pipeline init --template classifier-pipeline-resnet	clarifai pipeline run
"Train object detector"	clarifai pipeline init --template detector-pipeline-yolof	clarifai pipeline run
"Test model locally"	clarifai model serve ./model	Always try first
"Test as gRPC server"	clarifai model serve ./model --grpc --port 8000
"Test in container"	clarifai model serve ./model --mode container
"Deploy model to cloud"	clarifai model deploy ./model	Auto-creates infra
"Deploy with specific GPU"	clarifai model deploy ./model --instance gpu-nvidia-a10g
"Deploy already-uploaded model"	clarifai model deploy --model-url <url>
"Check deployment status"	clarifai model status --deployment <id>	Use --model-url to discover deployments
"View deployment logs"	clarifai model logs --deployment <id>	Follows by default
"View K8s events"	clarifai model logs --deployment <id> --log-type events	For debugging startup
"Remove deployment"	clarifai model undeploy --deployment <id>
"Browse available GPUs"	clarifai list-instances	Filter: --cloud, --gpu

References

• Complete SDK + CLI reference: references/sdk-cli-reference.md