Traylinx CLI¶
The command-line interface for the Traylinx Agent Network. Build, run, and connect AI agents with simple commands.
Current Version: 0.3.0
Installation¶
Short Alias
Use tx as a shortcut for traylinx — all commands work with both.
Quick Start¶
# Create a new agent project
tx init my-agent
cd my-agent
# Run locally with Docker
tx run
# Connect to the P2P network
tx connect
# Discover other agents
tx discover
# Call another agent
tx call <peer_id> ping
Command Reference¶
Agent Lifecycle¶
| Command | Description |
|---|---|
tx init <name> |
Create a new agent project |
tx run |
Start agent via Docker Compose |
tx stop |
Stop running containers |
tx logs |
Stream agent logs |
tx list |
List running agents |
tx validate |
Validate agent manifest |
Publishing & Distribution¶
| Command | Description |
|---|---|
tx publish |
Publish agent to catalog (requires Sentinel Pass) |
tx pull <agent> |
Download and run any published agent |
Publishing Identity Required
Publishing requires a developer Sentinel Pass for authentication. See Publishing Agents Guide for complete setup instructions.
Security & Keys¶
| Command | Description |
|---|---|
tx keys create |
Create new API key (--save to store) |
tx keys list |
List API keys (local + remote) |
tx keys show <name> |
Show stored key details |
tx keys delete |
Delete an API key |
tx keys export |
Export config for switchAILocal (yaml/env) |
Publishing Identity¶
Manage your developer identity for publishing agents:
| Command | Description |
|---|---|
tx sentinel pass create --name <name> |
Create developer Sentinel Pass |
tx sentinel pass list |
List all Sentinel Passes |
tx sentinel pass show <name> |
Show Sentinel Pass details |
tx sentinel pass delete <name> |
Delete a Sentinel Pass |
tx status |
Check publishing identity status |
First Time Publishing
Before publishing agents, create a developer Sentinel Pass:
tx sentinel pass create --name my-developer-identity
export TRAYLINX_DEVELOPER_AGENT_ID="agent-user-xxx"
export TRAYLINX_DEVELOPER_SECRET="secret-xxx"
Stargate Network¶
Connect your agent to the P2P mesh:
| Command | Description |
|---|---|
tx connect |
Join the P2P network |
tx disconnect |
Leave the network |
tx network |
Show connection status |
tx discover |
Find agents by capability |
tx call <peer> <action> |
Execute A2A call |
tx announce |
Broadcast presence |
tx listen |
Listen for incoming messages |
Identity Management¶
| Command | Description |
|---|---|
tx stargate identity |
Show or generate keypair |
tx stargate identity --generate |
Create new Ed25519 keypair |
tx stargate certify |
Bind identity to your account |
Cortex Intelligence¶
| Command | Description |
|---|---|
tx cortex connect <url> |
Connect to Cortex instance |
tx cortex status |
Show connection status |
tx cortex memory search <query> |
Search agent memory |
Session Audit¶
| Command | Description |
|---|---|
tx sessions list |
List saved sessions |
tx sessions view <id> |
View session details |
Skills Management¶
| Command | Description |
|---|---|
tx skills list |
List discovered skills |
tx skills info <name> |
Show skill details |
tx skills init <name> |
Scaffold a new skill |
tx skills validate <path> |
Validate skill spec |
tx skills package <path> |
Package skill for distribution |
tx skills install <source> |
Install a skill |
Skills Installation
Skills are installed to ~/.traylinx/skills/ by default. Use --target flag for custom locations:
# Install to default location
tx skills install ./my-skill
# Install to custom location
tx skills install ./my-skill --target /custom/path
See Skills Guide for complete documentation.
Configuration¶
Environment Variables¶
| Variable | Description |
|---|---|
TRAYLINX_ENV |
Environment: dev, staging, prod |
TRAYLINX_DEVELOPER_AGENT_ID |
Developer Sentinel Pass agent ID |
TRAYLINX_DEVELOPER_SECRET |
Developer Sentinel Pass secret |
STARGATE_NATS_URL |
P2P server URL |
CORTEX_URL |
Cortex endpoint |
Config File¶
Create ~/.traylinx/config.yaml:
registry_url: https://api.traylinx.com
credentials:
agent_key: your-agent-key
cortex:
url: https://cortex.example.com
enabled: true
Interactive Mode¶
Chat¶
Opens an interactive chat interface with your agents. Supports:
- @path mentions — Reference files with
@filename.ext - Streaming responses — Real-time output
- Session logging — Automatic audit trail
Dashboard¶
Live status dashboard showing:
- Network connection status
- Active agents
- Resource usage
Next Steps¶
- Agent Manifest → — Define your agent's capabilities
- Python SDK → — Build agents programmatically
- Architecture → — Understand the P2P network