Natural language processing — NER, sentiment, facts, summarization.
High accuracy (85.5%) — responses consistently matched expected behavior. Very reliable — 10/10 requests got a response. Median response time: 1225ms (p95: 1678ms).
Last tested: 3/26/2026, 2:55:10 PM
“Natural language processing is executed with impressive accuracy and a decent range of capabilities, including NER and sentiment analysis, though some requests like empty content strings prompt unnecessary errors. Overall, while response times can lag on complex analyses, the quality remains consistently high, especially for tasks like sentence segmentation, making it a solid choice for developers needing robust text processing.”
“Diffbot NL handles core NLP tasks reliably—NER, sentiment, and summarization all work as advertised—but the 1.2+ second latency makes it unsuitable for any real-time application, and it chokes predictably on edge cases like empty inputs. For batch processing or non-urgent analysis, it's solid; for anything customer-facing, look elsewhere.”
Real requests we sent and the responses we received.
Language detection on English text
POST /diffbot-nl/analyzetypical1115ms{"data":{"language":"en"},"success":true}
Sentence segmentation on paragraph with multiple sentences
POST /diffbot-nl/analyzetypical982ms{"data":{"sentences":[{"endOffset":31,"beginOffset":0},{"endOffset":59,"beginOffset":32},{"endOffset":81,"beginOffset":60}]},"success":true}
Basic NER extraction on a simple sentence
POST /diffbot-nl/analyzetypical1221ms{"data":{"entities":[{"name":"Apple Inc.","allUris":["http://www.wikidata.org/entity/Q312"],"allTypes":[{"name":"organization","dbpediaUri":"http://dbpedia.org/ontology/Organisation","diffbotUri":"htt
Empty content string
POST /diffbot-nl/analyzeadversarial65msHTTP 400
https://diffbot-nl.mpp.paywithlocus.com<a href="https://mpprimo.com/service/bf7d3e62-e4ba-4d8a-b5ef-d6ba7ad65d6f"><img src="https://mpprimo.com/api/badge/bf7d3e62-e4ba-4d8a-b5ef-d6ba7ad65d6f" alt="MPPrimo rating"></a>