​
Installation

pip install authed

Parameters in [square brackets] are optional. All other parameters are required.

​
Global configuration

The fastest way to get started is using our setup command:

authed init setup --provider-id "your-provider-id" --provider-secret "your-provider-secret"

This command will:

  • Generate a secure key pair for your agent
  • Create a new agent under your provider
  • Set up your environment variables automatically

​
Commands

​
Initialize and Setup

Complete setup (recommended):

authed init setup --provider-id <PROVIDER_ID> --provider-secret <PROVIDER_SECRET>

Show current config:

authed init show

Clear config (—force skips confirmation):

authed init clear [--force]

​
Agent management

List agents:

authed agents list [--output <FILE>]

Create agent:

authed agents create --name <AGENT_NAME> [--key-file <KEY_FILE>]

Get agent info:

authed agents info <AGENT_ID>

Delete agent (—force skips confirmation):

authed agents delete <AGENT_ID> [--force]

​
Key management

Generate keypair:

authed keys generate [--output <FILE>]

Show public key:

authed keys show-public <KEY_FILE>

Show private key (—force required as safety measure):

authed keys show-private <KEY_FILE> --force

​
Permission management

List permissions:

authed permissions list <AGENT_ID> [--output <FILE>]

Grant agent permission:

authed permissions grant <AGENT_ID> \
  --type allow_agent \
  --target-id <TARGET_AGENT_ID>

Grant provider permission:

authed permissions grant <AGENT_ID> \
  --type allow_provider \
  --target-id <TARGET_PROVIDER_ID>

Revoke permission (—force skips confirmation):

authed permissions revoke <AGENT_ID> \
  --type allow_agent \
  --target-id <TARGET_ID> \
  [--force]

Clear permissions (—force skips confirmation):

authed permissions clear <AGENT_ID> [--force]

​
Log management

Stream logs:

authed logs stream \
  --agent-id <AGENT_ID> \
  [--level INFO|WARNING|ERROR] \
  [--output <FILE>]

Fetch logs:

authed logs fetch \
  --agent-id <AGENT_ID> \
  [--level INFO|WARNING|ERROR] \
  [--limit <NUMBER>] \
  [--output <FILE>]

​
Common workflows

​
Setting up a new agent

  1. Generate keys:
authed keys generate --output agent_keys.json
  1. Create agent:
authed agents create \
  --name "my-agent" \
  --key-file agent_keys.json
  1. Verify creation:
authed agents info <AGENT_ID>

​
Managing permissions

List current permissions:

authed permissions list <AGENT_ID>

Grant agent access:

authed permissions grant <AGENT_ID> \
  --type allow_agent \
  --target-id <TARGET_AGENT_ID>

Grant provider access:

authed permissions grant <AGENT_ID> \
  --type allow_provider \
  --target-id <TARGET_PROVIDER_ID>

​
Output formats

​
JSON output

When using --output flag, data is saved in JSON format:

{
  "agents": [
    {
      "agent_id": "...",
      "name": "...",
      "provider_id": "...",
      "created_at": "...",
      "permissions": [
        {
          "type": "allow_agent",
          "target_id": "..."
        }
      ]
    }
  ]
}

​
Terminal output

Without --output, the CLI provides colored, formatted output:

  • Success messages in green (✓)
  • Warnings in yellow (⚠️)
  • Errors in red (✗)
  • IDs and technical values in different colors for readability