Phidata CLI Cheatsheet - phi Multi-agent Workspace Full Reference

Delivering an agent application often costs more in environment drift than in the logic itself, because every new machine, teammate, or staging box means reconfiguring dependencies and credentials again. Phidata, now part of the Agno ecosystem, takes the opposite stance: the phi command line initializes the framework, creates and switches workspaces, and turns resource declarations into a local Docker environment or real resources on AWS. This table groups initialization and auth, workspace management, and resource deployment, focusing on how configuration files map to environments and when the local docker profile should give way to the production aws one. The test is blunt: can a clean machine restore the entire environment with one command, or do you still have to hand-write an .env? Reach for it when several people develop an agent together, when a demo must become a shippable service, or when you want credentials out of every individual workstation. Against a hand-rolled service plus Dockerfile, the extra abstraction is resource declarations plus deployment targets, which pays off only if you genuinely run multiple environments.

AI CLI·21 commands·Last updated 2026-09-10
phidataphiaiagentworkspaceclideployment

Init and Auth 5

pip install -U phidata
Install or upgrade Phidata
phi init
Initialize the framework (creates ~/.phi)
phi auth
Web OAuth auth with phidata.com
phi config
Show current config and workspaces
phi --help
List all commands

Workspace Management 7

phi ws create -t agent-app -n my-app
Create a workspace from a template
phi ws setup
Set up an existing workspace
phi ws up
Start workspace resources locally
phi ws down
Stop and clean workspace resources
phi ws list
List available workspaces
phi ws up --group app
Start only the app resource group
phi ws up dev:docker
Start with dev/docker env

Resources and Deploy 9

phi start
Deploy resources from a file
phi stop
Stop resources from a file
phi set active-ws my-app
Set the active workspace
phi ws up prd:aws
Deploy to AWS production
phi ws up --env dev --infra docker
Filter deploy by env and infra
phi ping
Test auth status
phi reset
Reset the phi install
phi ws patch
Update workspace resources and config
phi login
CLI login with email and password

Tips

  • phi ws create -t starts from official templates; begin with agent-app or llm-app for the fastest path.
  • Use phi ws up and phi ws down as a pair: Docker locally, prd:aws for production.
  • Data lives in your own VPC or AWS account, matching bring-your-own-cloud compliance.

FAQ

How is phi different from crewai?

phi leans toward app plus infrastructure deployment (AWS/Docker orchestration); crewai is a multi-agent framework. The former handles delivery and ops, the latter agent orchestration.

Must I deploy to AWS?

No. phi ws up uses Docker locally; AWS is an optional production target, and data can stay fully in your own environment.

Official References

Each command links to its official documentation below, so you can verify the latest usage and read deeper.

Maintained by LaoHand

Publicly updated on Sep 10, 2026, continuously proofread against official docs.

Contact Us

Wrong command or description? Send us corrections, business inquiries or product feedback by email.

Contact Us