Batch your requests

Instead of making 10 calls with count=1, make one call with count=10. You get the same result with 1/10th the rate limit usage. Each API call counts as one request regardless of how many words it returns.

Cache the daily word

The daily word is identical for all users in the same timezone and changes once per day. Cache it aggressively — there's no benefit to calling it more than once per day. Use Cache-Control: max-age=86400 or your framework's caching (e.g. Next.js revalidate).

Use rate limit headers proactively

Every response includes these headers:

• X-RateLimit-Limit — your daily cap (e.g. 100 for free, 5000 for plus)
• X-RateLimit-Remaining — requests left today
• X-RateLimit-Reset — Unix timestamp when the rolling 24h window resets
• X-Burst-Limit / X-Burst-Remaining — per-minute burst counters

Check X-RateLimit-Remaining before making a request in tight loops to avoid hitting the 429 wall.

Keep your API key secret

Never put your API key in client-side JavaScript, public repositories, or URL query parameters. Use a server-side proxy (like a Next.js API route or Express endpoint) to call the Anagrama API from the server, and expose a keyless endpoint to your frontend. See the React Widget example for a complete pattern.

Use wordsets for custom word lists

If you need domain-specific words (e.g. animals, food, science terms), create a wordset instead of filtering random words client-side. Wordsets support hints, categories, and server-side scrambling — everything you need for vocabulary quizzes, flashcards, and custom puzzles.