Your first deploy

This guide takes your code to a running Service, whichever way it arrives. Your AI assistant can do every step for you. The rest of this page follows the same steps in Tokay's web interface.

Last updated

AI agent? Start with llms.txt.

Bring the code into Tokay

Your AI assistant. Connect your assistant to Tokay once and ask it to deploy. Your assistant sends the code, Tokay sets it up and checks that it runs, and you're asked only for what needs you. Use your AI assistant walks through the setup.

GitHub. Open Add code and choose GitHub, then pick the repository. Tokay copies and analyzes it, and GitHub stays where you commit. Deploy from GitHub covers the connection.

Paste or upload. For code that isn't in a repository yet, open Add code and choose Paste or upload. The editor accepts pasted code, a dragged folder, or one or more selected files, and shows a file sidebar when there are several. Choose Analyze when the files look right. When a safe deployment needs a small code change, such as moving a hardcoded API key into an environment variable, Tokay shows the proposed diff and the reason. Choose Apply changes, keep editing, or use your code as it is.

Git. Open Add code and choose Git to create a Tokay hosted repository, then push to it. Deploy from Git covers credentials and commands.

However the code arrives, Tokay identifies what it is, how it starts, which Service type it fits, and what values or resources it expects.

Review the detected setup

The configuration page shows the Service name, type, and stack detected in the code, along with the branch and the Project the Service goes into. In a new workspace that is a new Project called Default Project, which you can rename before going live. For a scheduled job, set the schedule in plain language and confirm the displayed upcoming runs.

If Tokay does not find a runnable Service, add or clarify an entry point such as server.js, app.py, or index.html, then push or analyze the code again.

Deploy and verify the result

Choose Go live, or Test & schedule for a scheduled job. The tracker moves through Code, Review, Build, and Go live. A first build usually takes a minute or two, while later updates are often faster.

If the code reads values Tokay cannot supply, such as an API key, the Service page shows a Finish setup panel with a field for each one. Values are saved as Secrets that nobody can read back. Choose Keep it readable only for a value that people and agents with source access may see, then choose Save and go live. Config, secrets, and environment variables explains how those values are shared across the Project.

When it finishes, verify the result for the Service type.

  • Open the URL for a web app.
  • Use Run now and inspect the output for a scheduled job.
  • Copy the URL and send a test request for a webhook.
  • Check the live logs for a background worker.

The Service page is the operational home for later deployments, logs, environment variables, and settings. See Editing your code for updates and Logs and debugging when the first run does not behave as expected.