Google Translate vs. Grammarly: data protection per plan

not publicly documented (consumer service under the Google Terms of Service; in Google Workspace, too, Google Translate is not a Core Service but an additional service for which Workspace Help names the Google Terms of Service, q9; no DPA is mentioned)

yes (data usage FAQ: 'Cloud Translation API complies with the Cloud Data Processing Addendum', q2, q4)

yes (data usage FAQ: 'Cloud Translation API complies with the Cloud Data Processing Addendum'; according to the FAQ, glossaries are also stored in accordance with the CDPA, q2, q4)

no for individual accounts (terms and conditions 3.5: Privacy Policy applies); if the account is opened on behalf of an organisation, the Data Privacy Addendum applies

no for individual accounts (terms and conditions 3.5: Privacy Policy applies); if the account is opened on behalf of an organisation, the Data Privacy Addendum applies

yes (automatically for accounts on behalf of an organisation; terms and conditions 3.5)

yes (Customer Business Agreement 3.2: DPA "incorporated by reference")

A Translate-specific statement on the use of entered texts for model training and an opt-out are not documented. The Google privacy policy (effective from 2026-10-16) states: 'We analyze usage of Google Translate to improve translation quality and increase the availability of Translate in more languages.' as well as the analysis of how often translations are marked as poor quality; publicly available information is used to train models for products such as Google Translate (q6).

Data usage FAQ (as of 2026-10-13): 'Google does not use any of your content for any purpose except to provide you with the Cloud Translation API service.' and 'Google does not use the content you send to train and improve our Google Translation features.' Content is not shared with third parties (q2).

Data usage FAQ (as of 2026-10-13): 'Google does not use any of your content for any purpose except to provide you with the Cloud Translation API service.' and 'Google does not use the content you send to train and improve our Google Translation features.' (q2). Custom models are trained with the customer's training data in the customer's project (pricing page: training billed by the hour, q1).

Product Improvement and Training is on by default for individual users; can be switched off in the account. Help Center: 'Product Improvement and Training is on for individual users by default.' Privacy policy: 'We also use information we collect to train our AI models.' According to the privacy policy, AI service providers may not train on customers' user content.

The same control applies to Free, Premium and single-user Pro, default on. Help Center: 'allows users ... to opt out of allowing us to use their content ... to train our models'. Non-content data (writing statistics) may continue to be used.

For accounts with more than one user, only the administrator changes the setting; default 'is enabled' (on). Help Center: 'Only admins can change the Product Improvement and Training setting for multi-user accounts.' Note: the Pro pricing page no longer lists 'Business' as a separate plan; the Help Center still refers to 'Grammarly Business'.

Help Center: 'For all Enterprise-tier accounts ... Product Improvement and Training is off by default' (also Pro/Business accounts purchased via Sales and all Education accounts). Customer Business Agreement Section 5.4: 'will not use Customer Data to train any generative artificial intelligence models without Customer's prior written consent.' The DPA itself contains no training clause. Privacy policy: AI service providers may not train on user content.

Google privacy policy: 'We maintain servers around the world and your information may be processed on servers located outside of the country where you live.' A Translate-specific location statement or region selection is not documented (q6).

Data usage FAQ: Cloud Translation – Basic has global endpoints (regional endpoints are documented only for Advanced, q3); 'the Services are global and data cannot be configured to remain in a specific region for processing' (q2). CDPA: subject to agreed location commitments, customer data may be processed in any country in which Google or its sub-processors maintain facilities (q4).

Documentation 'Global and multi-regional endpoints' (as of 2026-10-13): besides the global endpoint there are translate-eu.googleapis.com and translate-us.googleapis.com; with a multi-regional endpoint, data at rest and ML processing stay within the EU or the US respectively. The default is the global endpoint. Custom AutoML models and Preview features are not supported on the multi-regional endpoints (only the pre-trained NMT model). Organisation policy administrators can restrict the regions available for Advanced resources via a resource locations constraint (q3). The data usage FAQ refers to the 'AI/ML Data Location' section of the Service Specific Terms for regional endpoints (q2).

Security page: Grammarly 'hosts data in Amazon Web Services data centers in the US East region'. DPA Annex B 1.1: hosting on AWS and/or GCP, 'located in the United States'; Annex B 2.1: 'Grammarly is hosted on Amazon Web Services (AWS)' (the replication 'to a second geographical region' in 2.3 concerns only the Docs service); DPA 9: processing 'in the United States and any other locations'. Sub-processor list: all third-party providers with country USA; internal group companies including in Germany, Poland, Ukraine, Canada. An EU hosting option is not documented.

See Free.

See Free.

See Free (DPA Annex B 1.1: hosting in the USA).

• not publicly documented (consumer service, no list)

• Google Cloud sub-processor list: third parties mainly for technical support (e.g. Accenture, Infosys, NTT DATA, TELUS, Virtusa; for all Google Cloud Platform services) and for data labelling in AI services/Pre-Trained APIs (e.g. Cognizant, GlobalLogic, Innodata), each with countries, as well as Google group companies; the list contains no Translation-specific entries (q5)

• Google Cloud sub-processor list: third parties mainly for technical support (e.g. Accenture, Infosys, NTT DATA, TELUS, Virtusa; for all Google Cloud Platform services) and for data labelling in AI services/Pre-Trained APIs (e.g. Cognizant, GlobalLogic, Innodata), each with countries, as well as Google group companies; the list contains no Translation-specific entries (q5)

• Anthropic (Generative AI)
• AWS (Hosting)
• Azure (Generative AI)
• Baseten
• CoreWeave
• Databricks
• Deepgram (transcription)
• Google (infrastructure)
• OpenAI (Generative AI)
• Soniox (transcription)
• Turbopuffer (infrastructure)
• others as per the list (including Zendesk, Twilio, Sumo Logic); internal group companies in the USA, Germany, Poland, Ukraine, Canada

• see Free

• see Free

• see Free

0 (free service; a paid consumer plan is not documented on the retrieved pages)

In USD (pricing page): NMT text translation and language detection free for the first 500,000 characters per month (as a credit of USD 10 per month, shared between Basic and Advanced, does not roll over, not for formatted document translation); 500,000 to 1 billion characters per month USD 20 per 1 million characters (with Gemini Enterprise Flexible Savings Plan 1 or 3 years: USD 18 or 16); over 1 billion characters per month price via sales. Billed monthly by characters sent. The pricing page contains no information on taxes (q1).

In USD per 1 million characters (pricing page; discounts with Gemini Enterprise Flexible Savings Plan 1/3 years in brackets): NMT text free for the first 500,000 characters per month (USD 10 credit, shared with Basic), then USD 20 (18 / 16); NMT document translation USD 0.08 per page (0.072 / 0.064); custom models text tiered by monthly volume USD 80 up to 250 million, USD 60 up to 2.5 billion, USD 40 up to 4 billion, USD 30 above (with discounts down to USD 24); custom models document translation USD 0.25 per page (0.225 / 0.20); Adaptive Translation USD 25 input and USD 25 output (22.50 / 20); Translation LLM text USD 10 input and USD 10 output; Custom Text Translation (LLM) USD 10 input and USD 10 output (9 / 8). Training custom models USD 45 per hour, at most USD 300 per training job. The pricing page contains no information on taxes (q1).

Free: 0 EUR/month; 100 AI prompts/month (as of 2026-10-14, pricing page)

Pro: 12 EUR per member/month with annual billing (only '€12 EUR' visible; the plan data embedded in the pricing page source code lists 'Annual Grammarly Pro plan' 144.0 EUR, 'Quarterly' 60.0 EUR and 'Monthly Grammarly Pro plan' 30.0 EUR; not stated whether incl. VAT; 7-day trial); 2,000 AI prompts per member/month (as of 2026-10-18)

Pro price see above; payment by invoice for 10 or more Business members on the annual plan on request (pricing page FAQ)

Enterprise: price on request ('Contact Sales'); unlimited AI prompts per member/month (as of 2026-10-14)

yes (machine translation of words, phrases and web pages; Workspace Help: 'Instantly translate words, phrases, and web pages between English and over 100 other languages', q9)

yes (Neural Machine Translation (NMT): language detection and text translation, q1)

yes (NMT model, Translation LLM, Adaptive Translation, custom models (AutoML Translation), batch and document translation (DOCX, PPT, PDF, XLSX), romanisation, glossaries, q1)

yes

yes

yes

yes