|Docs

railway agent

Interact with the Railway Agent directly from the CLI. Ask questions about your project, create resources, debug issues, and more using natural language. Starts an interactive session by default, or send a single prompt with the -p flag.

Usage

Options

FlagDescription
-p, --prompt <MESSAGE>Send a single prompt (omit for interactive mode)
--jsonOutput response as JSON
--thread-id <ID>Continue an existing chat thread
-s, --service <SERVICE>Service to scope the conversation to (name or ID)
-e, --environment <ENV>Environment to use (defaults to linked environment)

Examples

Start an interactive session

Show logs for your project

Create a database

Debug a failing service

Check deployment status

Investigate resource usage

Scope to a specific service

Get configuration help

JSON output for scripting

Continue a previous conversation

Interactive mode

When run without -p, the agent starts an interactive session where you can have a back-and-forth conversation:

The agent maintains context within the session, so follow-up questions understand previous messages.

JSON output

Use --json to get a structured response for scripting and automation:

Returns a JSON object with:

Authentication

The agent command requires user authentication via railway login. Project access tokens (RAILWAY_TOKEN) are not supported.