databricks-mason

Mason is an experimental CLI for Databricks custom agent preview APIs and deployments. It manages memory, sessions, tracing, and deployments from one authenticated command.

The underlying APIs are in preview and may need workspace enablement.

Installation

From PyPI:

pip install databricks-mason

From source:

pip install 'git+https://github.com/databricks/databricks-ai-bridge.git#subdirectory=integrations/mason'

For tracing commands, install Mason with tracing extras:

pip install 'databricks-mason[tracing]'

Authentication

Mason uses Databricks authentication. If you do not already have credentials, authenticate a named profile first. You can then ask Mason to validate and remember that profile:

databricks auth login --profile <profile>
mason login --profile <profile>
mason sessions stores list

mason login does not create credentials; it stores the selected profile in ~/.mason/config.json. mason logout forgets that selection without revoking the underlying credentials. If Databricks SDK default authentication is already configured, you can skip mason login. You can also pass --profile/-p for an individual command. Use --output json for scripting.

Commands

mason [-p <profile>] [-o text|json]
  login        [--profile P]
  logout
  memory
    stores     create | list | get | update | delete
    entries    create | get | list | search | update | delete
  sessions     create | list | get | update | delete | fork
    stores     create | list | get | update | delete
    items      list | append | pop | clear
  tracing
    setup      --catalog C --schema S [--experiment E]
    list | get | instrument
  deploy       <name> --source PATH [--with-memory-store N]
               [--with-session-store N] [--with-traces C.S] [--create-stores]
  deployments  list | get | logs | start | stop | delete