AI File Namer

A small SvelteKit app that uses Google Gemini (via @google/genai) to suggest well-formatted filenames from a short description. It exposes a secure server endpoint that calls the GenAI model and returns a JSON array of suggested filenames.

This repository is a compact demo and starter for integrating a text-generation model into a web UI where the model output is constrained and validated.

Features

• SvelteKit frontend with a textarea to enter a document description.
• Server-side endpoint (/api/generate) that calls the GenAI client and returns 5 filename suggestions.
• Input validation using zod (trim, non-empty, min/max length).
• Clipboard-friendly UI with a smooth "Copy → Copied" transition for suggested names.
• Defensive server parsing and helpful dev-mode error messages.

Stack / Key dependencies

• SvelteKit
• Vite
• @google/genai (Gemini client)
• zod (input validation)

See package.json for exact versions.

Prerequisites

• Node.js 18+ (or the version your environment expects)
• npm (or pnpm/yarn)
• A Google Gemini API key with access to the model used in this project. Store it in an environment variable called GEMINI_API_KEY.

Setup

1. Clone the repository and install dependencies (from project root):

npm install

2. Create a .env.local file at the project root (next to package.json) and add your key:

GEMINI_API_KEY=YOUR_REAL_KEY_HERE

Important: Do NOT commit .env.local to source control. Ensure your .gitignore contains it.

3. Start the dev server:

npm run dev

Open http://localhost:5173 (or the port printed by dev server) and try the textarea.

API: /api/generate

• Method: POST
• Content-Type: application/json
• Body shape: { "description": "...your description..." }

Success response:

{ "success": true, "names": ["file-name-1.pdf", "file-name-2.docx", ...] }

Failure responses include a success: false and an error message. Validation errors return HTTP 400 with a message assembled from Zod issues.

Example curl

curl -X POST 'http://localhost:5173/api/generate' \
    -H 'Content-Type: application/json' \
    -d '{"description":"List of employee names in my company"}'

Validation rules (server-side)

Validation is implemented with src/lib/schema.js using zod:

• description is trimmed
• must be non-empty (custom message)
• must be at least 10 characters
• must be at most 500 characters

The server uses FilenameSchema.safeParse(body) to validate the incoming JSON before calling the model.

UI details

• src/routes/+page.svelte contains the main UI. It posts to the server route at api/generate and renders results.
• Copy buttons use the Clipboard API and show a transient "Copied" state for UX clarity.

Error handling & debugging

• The server route performs defensive parsing of the GenAI SDK response and returns helpful error messages in dev mode.
• If you get an error about the API key, ensure GEMINI_API_KEY is set and the dev server was restarted after creating/updating .env.local.
• If validation fails, check the description length and whitespace (the schema trims input first).

Deployment

• Use your hosting provider's secret management to set GEMINI_API_KEY in production instead of .env.local.
• This project includes @sveltejs/adapter-* packages in devDependencies; choose and configure an adapter for your target platform (Vercel, Netlify, Cloudflare Pages, etc.).

Security & privacy

• Never commit secrets. Use the hosting platform's environment variables for production.
• Model requests may be logged by the model provider—avoid sending highly sensitive personal or proprietary data unless your contract and security posture permit it.

Development notes & suggestions

• Convert server routes to TypeScript (+server.ts) if you want stronger types for request/response objects.
• Add tests that exercise src/lib/schema.js to assert validation behaviour across edge cases.
• Consider sanitizing or escaping user-supplied content before rendering or storing it.

Contributing

PRs are welcome. Keep changes small and focused. If adding features that change the API shape, update this README and add tests.

License

Choose a license for your project (e.g., MIT). This repo currently does not include a license file.