Abstract & Keyword Generator

There are 3 successful uses per day shared across abstracts and keywords, with 100 model attempts per day site-wide. Visitor and IP limits both apply; campus or office networks may share an allowance. Limits reset at midnight Beijing time.

Insufficient material, model failures and timeouts restore your free use, but still count toward attempt and site-wide limits. Wait at least 30 seconds between requests; the default daily attempt limit is 10.

No. Provide 200–100,000 characters of substantive source material. The tool summarizes existing content rather than inventing data, experiments or findings. Literary and other non-experimental research can use argument and analysis materials.

TXT, Markdown and DOCX text is extracted locally; the original file is not uploaded. After consent, text is sent via Vercel to DeepSeek. Source text is not stored in our database. Results are recoverable for 10 minutes, expired caches are purged hourly, and pseudonymous quota records are retained for 7 days. Providers process necessary data under their own terms.

Do not immediately resubmit. The page checks and recovers the original request without another model call. A validated server-side result counts as a successful use. Download or copy it before the recovery window expires.

Generated content is a draft and length settings are approximate targets. Verify facts, terminology, research boundaries and submission requirements, and follow your institution’s AI use and disclosure rules.