Original working file
START-HERE.txt
Code and working example pages keep their original identifiers and sample text. The lesson and file guidance explain them in your language.
PRACTICAL ROUTE
Use a supplied server-backed example to keep a provider key out of the visitor’s browser and verify allowed and rejected requests.
Your role: Provide account values privately, approve the intended setup and check the real browser/service result.
1. Use the supplied project and ask your coding assistant to carry out the implementation. You do not need to write HTML, JavaScript or SQL yourself.
2. Follow the current lesson steps, configure only your own permitted accounts and check the result in the actual page/provider.
3. Keep expected results and observed results separate. Record the exact blocked step if you cannot run a live check.
The detailed setup below is available to you and your assistant. Code explanations and extra implementation exercises are optional; important access/privacy checks are still part of the practical lesson.
--- REQUIRED SETUP (CODE READING IS OPTIONAL) ---
This is a separate practice project. public/ is the website folder; netlify/functions/ contains server-only functions. Install Node and Netlify CLI for local function testing, or deploy this project to your own Netlify account. Put real credentials in server environment variables, never in public/ or Git. No provider request runs just by opening these files. Account steps and real delivery must be checked in your own account.
Set GEMINI_API_KEY, GEMINI_MODEL (an available text model from your own AI Studio account) and an unguessable LAB_TOKEN at least 24 characters long. The practice code is entered by the owner, not bundled. Use synthetic messages only: they are sent to Google. Check current account pricing, quota and data terms first. This owner-restricted lab is not a public chat service; add user authentication, durable usage limits, monitoring and a privacy notice before public use. No hardcoded retired model or promised free unlimited tier.
This lesson sends one fixed question, not arbitrary visitor prompts. Deploy public/ plus netlify/functions/. Search public files and browser requests for the provider key before considering the move complete.
For exact local commands and private environment names, read run-locally.txt and server-env.example.txt.
To inspect current account-visible model IDs without hardcoding a retired name, run: node --env-file=.env list-text-models.mjs
This helper lists model metadata and prints names only; it does not request a generated answer. Pick a currently supported text candidate and check pricing/quota before setting GEMINI_MODEL. Current model guide: https://ai.google.dev/gemini-api/docs/models