The index
All ratings
Updated 10 Nov 2026 · 48 tools
13 Safe · 33 Careful · 2 Avoid
AllChat assistantsMeeting note takersVoice AICoding assistantsWriting toolsImage toolsBrowser extensionsAI searchWorkspace AI
ToolVendorCategoryVerdictLast checked
- Amazon Q BusinessCarefulAmazon Web Services · Workspace AI · 10 Nov 2026AWS says Amazon Q Business is a HIPAA-eligible service, and its data-protection page tells you to use IAM, MFA, TLS, and CloudTrail.Amazon Q BusinessAWS says Amazon Q Business is a HIPAA-eligible service, and its data-protection page tells you to use IAM, MFA, TLS, and CloudTrail.Amazon Web ServicesWorkspace AICareful10 Nov 2026
- AssemblyAICarefulAssemblyAI · Voice AI · 10 Nov 2026AssemblyAI's terms say customer data, including audio inputs, may be used to train its models, and that the opt-out applies only to the extent the customer's pricing plan allows it.AssemblyAIAssemblyAI's terms say customer data, including audio inputs, may be used to train its models, and that the opt-out applies only to the extent the customer's pricing plan allows it.AssemblyAIVoice AICareful10 Nov 2026
- Canva Magic StudioCarefulCanva · Image tools · 10 Nov 2026Canva's privacy policy explains account data, advertising choices, and a machine-learning use aimed at measuring traffic and usage.Canva Magic StudioCanva's privacy policy explains account data, advertising choices, and a machine-learning use aimed at measuring traffic and usage.CanvaImage toolsCareful10 Nov 2026
- CartesiaCarefulCartesia · Voice AI · 10 Nov 2026Cartesia's privacy policy, revised June 14, 2024, says content such as text prompts and voice or audio recordings is used to train and enhance the models, unless you submit an opt-out form.CartesiaCartesia's privacy policy, revised June 14, 2024, says content such as text prompts and voice or audio recordings is used to train and enhance the models, unless you submit an opt-out form.CartesiaVoice AICareful10 Nov 2026
- ChatGPTCarefulOpenAI · Chat assistants · 10 Nov 2026On personal ChatGPT plans (Free, Plus, and Pro), OpenAI says data sharing is on by default and that it may use your content to train models unless you turn it off.ChatGPTOn personal ChatGPT plans (Free, Plus, and Pro), OpenAI says data sharing is on by default and that it may use your content to train models unless you turn it off.OpenAIChat assistantsCareful10 Nov 2026
- ChatGPT BusinessCarefulOpenAI · Chat assistants · 10 Nov 2026OpenAI says ChatGPT Business workspace data is excluded from training by default, and each person has their own chat history.ChatGPT BusinessOpenAI says ChatGPT Business workspace data is excluded from training by default, and each person has their own chat history.OpenAIChat assistantsCareful10 Nov 2026
- ChatGPT EnterpriseSafeOpenAI · Chat assistants · 10 Nov 2026OpenAI says it does not train on ChatGPT Enterprise inputs or outputs by default, and workspace admins control how long data is kept.ChatGPT EnterpriseOpenAI says it does not train on ChatGPT Enterprise inputs or outputs by default, and workspace admins control how long data is kept.OpenAIChat assistantsSafe10 Nov 2026
- ClaudeCarefulAnthropic · Chat assistants · 10 Nov 2026Anthropic's consumer privacy center says chats and coding sessions on Free, Pro, and Max are used to improve models if you allow it, if a safety classifier flags them, or if you otherwise opt in.ClaudeAnthropic's consumer privacy center says chats and coding sessions on Free, Pro, and Max are used to improve models if you allow it, if a safety classifier flags them, or if you otherwise opt in.AnthropicChat assistantsCareful10 Nov 2026
- Claude for WorkSafeAnthropic · Chat assistants · 10 Nov 2026Anthropic's commercial terms say it may not train models on customer content, and the privacy center says commercial inputs and outputs are not used for training by default.Claude for WorkAnthropic's commercial terms say it may not train models on customer content, and the privacy center says commercial inputs and outputs are not used for training by default.AnthropicChat assistantsSafe10 Nov 2026
- Copy.aiCarefulCopyAI · Writing tools · 10 Nov 2026Copy.ai's privacy notice describes account data, email choices, cookies, and state privacy rights.Copy.aiCopy.ai's privacy notice describes account data, email choices, cookies, and state privacy rights.CopyAIWriting toolsCareful10 Nov 2026
- CursorCarefulAnysphere · Coding assistants · 10 Nov 2026Cursor says that with Privacy Mode on, your data is not used for training and model providers are under zero-data-retention agreements, with a carve-out if abuse detectors fire.CursorCursor says that with Privacy Mode on, your data is not used for training and model providers are under zero-data-retention agreements, with a carve-out if abuse detectors fire.AnysphereCoding assistantsCareful10 Nov 2026
- DeepgramCarefulDeepgram · Voice AI · 10 Nov 2026Deepgram's privacy notice, updated November 5, 2026, says it may use Content for AI model training unless you opt out.DeepgramDeepgram's privacy notice, updated November 5, 2026, says it may use Content for AI model training unless you opt out.DeepgramVoice AICareful10 Nov 2026
- DeepSeekAvoidDeepSeek · Chat assistants · 10 Nov 2026DeepSeek's privacy policy, updated February 10, 2026, names Hangzhou DeepSeek Artificial Intelligence Co., Ltd.DeepSeekDeepSeek's privacy policy, updated February 10, 2026, names Hangzhou DeepSeek Artificial Intelligence Co., Ltd.DeepSeekChat assistantsAvoid10 Nov 2026
- DescriptCarefulDescript · Writing tools · 10 Nov 2026Descript says it may use your projects to improve the service unless you turn off Share Data with Descript.DescriptDescript says it may use your projects to improve the service unless you turn off Share Data with Descript.DescriptWriting toolsCareful10 Nov 2026
- DevinCarefulCognition · Coding assistants · 10 Nov 2026Cognition's Devin security docs say that, by default, it may use your data for model training.DevinCognition's Devin security docs say that, by default, it may use your data for model training.CognitionCoding assistantsCareful10 Nov 2026
- ElevenLabsCarefulElevenLabs · Voice AI · 10 Nov 2026ElevenLabs says you can opt out of using your personal information and content for training, and that the opt-out only stops later training use.ElevenLabsElevenLabs says you can opt out of using your personal information and content for training, and that the opt-out only stops later training use.ElevenLabsVoice AICareful10 Nov 2026
- FathomCarefulFathom · Meeting note takers · 10 Nov 2026Fathom says third-party model providers are not allowed to train on your meetings, but Fathom itself uses de-identified meeting content to train its own models unless you opt out.FathomFathom says third-party model providers are not allowed to train on your meetings, but Fathom itself uses de-identified meeting content to train its own models unless you opt out.FathomMeeting note takersCareful10 Nov 2026
- FirefliesSafeFireflies.ai · Meeting note takers · 10 Nov 2026Fireflies says it does not use your meeting data to train AI models, and that its AI vendors do not retain that data after processing.FirefliesFireflies says it does not use your meeting data to train AI models, and that its AI vendors do not retain that data after processing.Fireflies.aiMeeting note takersSafe10 Nov 2026
- Gemini AppsCarefulGoogle · Chat assistants · 10 Nov 2026Google's Gemini Apps Privacy Hub says that when Keep Activity is on, chats can be used to train generative models and a subset can be read by human reviewers.Gemini AppsGoogle's Gemini Apps Privacy Hub says that when Keep Activity is on, chats can be used to train generative models and a subset can be read by human reviewers.GoogleChat assistantsCareful10 Nov 2026
- Gemini in Google WorkspaceSafeGoogle · Workspace AI · 10 Nov 2026Google says Workspace customer data, including prompts, is not used to train generative models outside your domain without permission, and Gemini does not share that content outside the organization without permission.Gemini in Google WorkspaceGoogle says Workspace customer data, including prompts, is not used to train generative models outside your domain without permission, and Gemini does not share that content outside the organization without permission.GoogleWorkspace AISafe10 Nov 2026
- Gemini NotebookCarefulGoogle · Workspace AI · 10 Nov 2026Google's Gemini Notebook help page, the product formerly called NotebookLM, says your data is not used to train Gemini Notebook unless you send feedback.Gemini NotebookGoogle's Gemini Notebook help page, the product formerly called NotebookLM, says your data is not used to train Gemini Notebook unless you send feedback.GoogleWorkspace AICareful10 Nov 2026
- GitHub Copilot Business and EnterpriseSafeGitHub · Coding assistants · 10 Nov 2026GitHub says it does not use Copilot Business or Copilot Enterprise data to train its models, and the 2026 individual-plan training change does not apply to those plans.GitHub Copilot Business and EnterpriseGitHub says it does not use Copilot Business or Copilot Enterprise data to train its models, and the 2026 individual-plan training change does not apply to those plans.GitHubCoding assistantsSafe10 Nov 2026
- GitHub Copilot Free, Pro, and Pro+CarefulGitHub · Coding assistants · 10 Nov 2026GitHub says that from April 24, 2026, interaction data from Copilot Free, Pro, and Pro+ is used to train and improve its models unless the user opts out.GitHub Copilot Free, Pro, and Pro+GitHub says that from April 24, 2026, interaction data from Copilot Free, Pro, and Pro+ is used to train and improve its models unless the user opts out.GitHubCoding assistantsCareful10 Nov 2026
- GrammarlyCarefulGrammarly · Browser extensions · 10 Nov 2026Grammarly's trust page says you can stop it from using your content to train models by turning off Product Improvement and Training.GrammarlyGrammarly's trust page says you can stop it from using your content to train models by turning off Product Improvement and Training.GrammarlyBrowser extensionsCareful10 Nov 2026
- GranolaCarefulGranola · Meeting note takers · 10 Nov 2026Granola says third parties such as OpenAI and Anthropic are not allowed to train on your data.GranolaGranola says third parties such as OpenAI and Anthropic are not allowed to train on your data.GranolaMeeting note takersCareful10 Nov 2026
- HumeCarefulHume AI · Voice AI · 10 Nov 2026Hume's public privacy policy, updated February 25, 2025, says it does not apply to content processed for business API customers.HumeHume's public privacy policy, updated February 25, 2025, says it does not apply to content processed for business API customers.Hume AIVoice AICareful10 Nov 2026
- JasperCarefulJasper · Writing tools · 10 Nov 2026Jasper's trust page says company data and outputs are never used to train third-party language models, and it lists SOC 2 among its compliance claims, plus SSO, audit logs, and permissions.JasperJasper's trust page says company data and outputs are never used to train third-party language models, and it lists SOC 2 among its compliance claims, plus SSO, audit logs, and permissions.JasperWriting toolsCareful10 Nov 2026
- KrispCarefulKrisp · Voice AI · 10 Nov 2026Krisp says noise cancellation alone does not send audio off the device, and that Krisp does not store or access those conversations.KrispKrisp says noise cancellation alone does not send audio off the device, and that Krisp does not store or access those conversations.KrispVoice AICareful10 Nov 2026
- Microsoft 365 CopilotSafeMicrosoft · Workspace AI · 10 Nov 2026Microsoft says prompts, responses, and data accessed through Microsoft Graph are not used to train foundation models for Microsoft 365 Copilot.Microsoft 365 CopilotMicrosoft says prompts, responses, and data accessed through Microsoft Graph are not used to train foundation models for Microsoft 365 Copilot.MicrosoftWorkspace AISafe10 Nov 2026
- Microsoft CopilotCarefulMicrosoft · Chat assistants · 10 Nov 2026Microsoft's consumer Copilot privacy FAQ says conversation activity can be used for model training and that you can opt out.Microsoft CopilotMicrosoft's consumer Copilot privacy FAQ says conversation activity can be used for model training and that you can opt out.MicrosoftChat assistantsCareful10 Nov 2026
- MurfCarefulMurf · Voice AI · 10 Nov 2026Murf's terms require explicit written consent from a person before their recordings are used to train a cloning voice, and they say Murf staff may listen to samples to test quality.MurfMurf's terms require explicit written consent from a person before their recordings are used to train a cloning voice, and they say Murf staff may listen to samples to test quality.MurfVoice AICareful10 Nov 2026
- Notion AISafeNotion · Workspace AI · 10 Nov 2026Notion says customer data is not used to train models by default, and its security page says customer data is never used to train models by Notion or its subprocessors.Notion AINotion says customer data is not used to train models by default, and its security page says customer data is never used to train models by Notion or its subprocessors.NotionWorkspace AISafe10 Nov 2026
- OtterAvoidOtter.ai · Meeting note takers · 10 Nov 2026Otter's privacy policy says it improves the service by training its own AI on de-identified audio recordings and on transcriptions that may still contain personal information.OtterOtter's privacy policy says it improves the service by training its own AI on de-identified audio recordings and on transcriptions that may still contain personal information.Otter.aiMeeting note takersAvoid10 Nov 2026
- PerplexityCarefulPerplexity · AI search · 10 Nov 2026Perplexity's help center says AI data retention is on by default for Free, Pro, and Max, and that this data may be used to train models unless you turn the toggle off.PerplexityPerplexity's help center says AI data retention is on by default for Free, Pro, and Max, and that this data may be used to train models unless you turn the toggle off.PerplexityAI searchCareful10 Nov 2026
- Perplexity EnterpriseSafePerplexity · AI search · 10 Nov 2026Perplexity's enterprise terms say it will not use customer content to train or fine-tune generative models, and its help center says the same promise binds third-party model providers.Perplexity EnterprisePerplexity's enterprise terms say it will not use customer content to train or fine-tune generative models, and its help center says the same promise binds third-party model providers.PerplexityAI searchSafe10 Nov 2026
- PlayAICarefulPlayAI · Voice AI · 10 Nov 2026PlayAI's terms say it will not train its models on content you submit, except aggregated usage metrics that do not reveal that content.PlayAIPlayAI's terms say it will not train its models on content you submit, except aggregated usage metrics that do not reveal that content.PlayAIVoice AICareful10 Nov 2026
- PoeCarefulQuora · Chat assistants · 10 Nov 2026Poe is Quora's chat product for talking to third-party model bots.PoePoe is Quora's chat product for talking to third-party model bots.QuoraChat assistantsCareful10 Nov 2026
- Read AISafeRead AI · Meeting note takers · 10 Nov 2026Read AI says contributing meeting data to model improvement is off unless you opt in, and that you can change your mind.Read AIRead AI says contributing meeting data to model improvement is off unless you opt in, and that you can change your mind.Read AIMeeting note takersSafe10 Nov 2026
- Resemble AISafeResemble AI · Voice AI · 10 Nov 2026Resemble says it does not use customer voice recordings to train general-purpose models, for research, or for aggregated training.Resemble AIResemble says it does not use customer voice recordings to train general-purpose models, for research, or for aggregated training.Resemble AIVoice AISafe10 Nov 2026
- RunwayCarefulRunway · Image tools · 10 Nov 2026Runway's privacy policy, served from runway.com, describes account data, analytics, payment processors, and state privacy opt-outs for advertising and sale.RunwayRunway's privacy policy, served from runway.com, describes account data, analytics, payment processors, and state privacy opt-outs for advertising and sale.RunwayImage toolsCareful10 Nov 2026
- Slack AISafeSlack · Workspace AI · 10 Nov 2026Slack says customer data never leaves Slack-controlled infrastructure and is never used to train large language models.Slack AISlack says customer data never leaves Slack-controlled infrastructure and is never used to train large language models.SlackWorkspace AISafe10 Nov 2026
- SpeechifyCarefulSpeechify · Voice AI · 10 Nov 2026Speechify's terms forbid using the service to make a deceptive voice clone without the explicit consent of the person whose voice is copied.SpeechifySpeechify's terms forbid using the service to make a deceptive voice clone without the explicit consent of the person whose voice is copied.SpeechifyVoice AICareful10 Nov 2026
- Stability AICarefulStability AI · Image tools · 10 Nov 2026Stability AI's privacy policy says it uses inputs and outputs to train, fine-tune, and evaluate models, and that you can opt out of that use in account settings.Stability AIStability AI's privacy policy says it uses inputs and outputs to train, fine-tune, and evaluate models, and that you can opt out of that use in account settings.Stability AIImage toolsCareful10 Nov 2026
- TabnineSafeTabnine · Coding assistants · 10 Nov 2026Tabnine's privacy docs say it has a no-train, no-retain policy for your code, including when other models are selected, and that code context is deleted after the suggestion is returned.TabnineTabnine's privacy docs say it has a no-train, no-retain policy for your code, including when other models are selected, and that code context is deleted after the suggestion is returned.TabnineCoding assistantsSafe10 Nov 2026
- tl;dvCarefulTldx Solutions · Meeting note takers · 10 Nov 2026tl;dv's privacy policy says it does not use meeting recordings, transcripts, notes, or files to train foundation or generative models, for itself or for third parties.tl;dvtl;dv's privacy policy says it does not use meeting recordings, transcripts, notes, or files to train foundation or generative models, for itself or for third parties.Tldx SolutionsMeeting note takersCareful10 Nov 2026
- WellSaidCarefulWellSaid Labs · Voice AI · 10 Nov 2026WellSaid's privacy policy says it may collect sensory data, including audio and voice, if you use video or audio capture in the service.WellSaidWellSaid's privacy policy says it may collect sensory data, including audio and voice, if you use video or audio capture in the service.WellSaid LabsVoice AICareful10 Nov 2026
- Wispr FlowCarefulWispr · Voice AI · 10 Nov 2026Wispr says it uses your audio, transcripts, and edits to train models only if you allow "Improve the model for everyone," and that any user can turn that off.Wispr FlowWispr says it uses your audio, transcripts, and edits to train models only if you allow "Improve the model for everyone," and that any user can turn that off.WisprVoice AICareful10 Nov 2026
- Zoom AI CompanionSafeZoom · Meeting note takers · 10 Nov 2026Zoom says it does not use audio, video, chat, screen sharing, attachments, or other communications-like customer content to train Zoom's or third-party AI models.Zoom AI CompanionZoom says it does not use audio, video, chat, screen sharing, attachments, or other communications-like customer content to train Zoom's or third-party AI models.ZoomMeeting note takersSafe10 Nov 2026