Storing, Caching & Retaining Output

Tavily’s Terms place few restrictions on what you do with the results you receive. You may store search result metadata, snippets, full page content, and complete API responses, including long-term, with no maximum retention period and no deletion requirement, even after your subscription ends. You may also cache responses, reuse stored output instead of making new API calls, keep it in internal exports and backups, and retain data you derive from it, such as summaries, extracted facts, and classifications. The same rules apply to Search, Extract, and Crawl results, and wherever results are stored.

Three points apply throughout. The underlying web content belongs to its publishers, so you’re responsible for respecting their rights and terms, particularly with full-page content. AI-generated answers carry a few extra conditions. And if you plan to share cached results across multiple customers of your own, contact support@tavily.com first so we can confirm the use fits your plan.

May I store search result metadata (URLs, titles, dates, rank, scores) in my own database long-term?

• Yes. Tavily's Terms do not restrict storing search result metadata (URLs, titles, dates, rank, scores) in your own systems, including long-term. The underlying web content belongs to its publishers, so you're responsible for respecting any rights or terms that apply to it.

May I store the snippets or content excerpts returned in search results?

• Yes. Snippets and excerpts are part of the results, and Tavily's Terms do not restrict storing them. The underlying content belongs to the original publishers, so your use of it must respect their rights and terms.

May I store full page content (raw_content, Extract output) or complete API responses?

• Yes. Tavily's Terms treat full page content and complete API responses the same as other results and does not restrict storing them. Full-page content is more likely to be protected by copyright, so make sure your use respects the publisher's rights and terms.

Is there a maximum retention period, or any deletion requirement, for stored output?

• No. Tavily's Terms do not set a maximum retention period, or require you to delete results you've stored.

Is temporary caching of API responses allowed, and for how long? (Includes caching shared across customers.)

• Yes. Tavily's Terms do noy limit caching of API responses or set a maximum cache period. If you plan to share cached results across multiple customers of your own, contact us at support@tavily.com first so we can confirm the use fits your plan.

Do I have to delete stored output when my subscription ends or my account closes?

• No. When your subscription ends you must stop using Tavily, but Tavily's Terms do not require you to delete results you've already stored. You will lose access to any data held in your Tavily account.

May I permanently keep only the source URL found through Tavily (e.g., as a provenance or citation record)?

• Yes. You can keep source URLs permanently, for example as a citation or provenance record.

If I visit a website found through Tavily and collect data from that site myself, does keeping that data count as storing Tavily output?

• No. Data you collect directly from a website yourself isn't Tavily output and isn't governed by Tavily's Terms. That website's own terms and applicable law apply.

May I keep data I derive from output, such as my own summaries, extracted facts, classifications, scores or content hashes?

• Yes. Tavily's Terms do not restrict data you creating from results summaries, extracted facts, classifications, scores or content hashes. You remain responsible for respecting the rights of the original content owners.

Do storage rules change if output is stored only on the end user's own device or browser?

• No. The same rules apply wherever results are stored, including on an end user's device or browser.

May I reuse stored output later (e.g., re-analysis, replay, showing it to other users) instead of making a new API call?

• Yes. You can reuse stored results, for example for re-analysis or replay, without making a new API call. Showing stored results to your application's end users is allowed on the same basis as live results.

May stored output be kept in internal exports (CSV/Excel) and backups?

• Yes. You may keep results in internal exports (such as CSV or Excel files) and backups.

Do storage and usage rights differ between Search results, Extract/raw content, Crawl, and generated answers?

• No. Tavily's Terms treat results from Search, Extract and Crawl the same way. AI-generated answers carry a few extra conditions: review them before relying on them, do not present them as human-written, and do not use them to build AI models that compete with Tavily.