Turn this comparison into procurement-ready diligence
Free shows today's risk. A Stack Audit gives you a citable, verbatim-sourced PDF across your whole AI stack — and flags the moment a vendor's terms change.
Free shows today's risk. A Stack Audit gives you a citable, verbatim-sourced PDF across your whole AI stack — and flags the moment a vendor's terms change.
Your privacy
We use first-party analytics and Google Analytics to understand traffic and improve the site (IP-derived location, device info, analytics cookies). We never sell your personal information and run no advertising trackers. You can opt out below. See our Privacy Policy and Cookie Policy.
293 verified findings across the displayed surfaces.
202 verified findings across the displayed surfaces.
Empty Comparison Slot
Cohere API carries highly permissive license terms; Perplexity AI carries highly permissive license terms.FOR ENTERPRISE: COHERE API NOT RECOMMENDED · PERPLEXITY AI NOT RECOMMENDED
| RISK SURFACE | COHERE API | PERPLEXITY AI |
|---|---|---|
| PPrompt Ownership | ||
| OOutput Ownership | ||
| TModel Training | ||
| CCommercial Use | ||
| DData Privacy | ||
| $Tier Differences |
“You agree that any suggestion or idea provided by you (such suggestions or ideas, “ Feedback ”) will not be treated as confidential, and nothing in these Terms of Use will restrict our right to use, profit from, disclose, publish or otherwise exploit any Feedback, without compens…”
Feedback submitted by users — which could be interpreted broadly to include prompts or suggestions provided to the platform — is subject to a perpetual, irrevocable, royalty-free worldwide license with no compensation to the user.
“ In order to operate the Service, we must obtain from you certain license rights in Your Content so that actions we take in operating the Service are not considered legal violations. Accordingly, by using the Service and uploading Your Content, you grant us a license to access, u…”
“ publish, market, advertise or in any way distribute the Content;”
“ Information we receive when you choose to link any third-party platforms to your account, such as when you sign into your account through Google or Apple (“ Third Party Platform Information ”) or when you voluntarily link one or more of your health and wellness applications (suc…”
“For Trial Users and Researchers (i.e. users interacting with Cohere Products on research-specific platforms like Hugging Face, Cohere Labs grantee users, or users demoing Cohere Products under a free trial account), we collect content you submit to the Cohere Products, and output…”
Trial Users and Researchers have their inputs and outputs used for model training and R&D by default, with no explicit opt-out described in this section.
“In addition to the specific uses described above, we may use most of the above information to provide you with and improve the Services (including our AI models) and to maintain our business relationship, including by enhancing the safety and security of our Services (e.g., troub…”
User data — which includes 'most' categories described above (including Service Interaction Information, i.e., prompts and outputs) — may be used to improve AI models. No opt-out mechanism is described here.
“ attempt to gain unauthorized access to the Cohere Solution, or bypass any measures we may use to prevent or restrict access to the Cohere Solution, attempt to circumvent the intended features, functionality or limitations of the Cohere Solution, or otherwise use, copy, distribut…”
“ If we become aware that we have unknowingly collected information about a child under 13 years of age or the relevant minimum age in your jurisdiction, we will make commercially reasonable efforts to delete such information. If you are the parent or guardian of a child under the…”
“Safeguards, Retention and Privacy Governance”
The document references a 'Safeguards, Retention and Privacy Governance' section but does not provide its content in the supplied text. Retention periods and deletion rights cannot be assessed from this excerpt.
“To the extent the laws in your jurisdiction do not recognize the legal basis of legitimate interest or another legal basis specified above for a particular purpose, you consent to the processing of your personal information for that purpose through using the Services.”
Perplexity uses mere use of the Services as implied consent for processing where 'legitimate interests' is not recognized as a valid legal basis. This is a broad fallback consent mechanism that may not be valid under GDPR or other strict privacy regimes.
“Section 2.3 (Research and Model Improvement): Analyzing usage patterns and using Trial or Research User Inputs/Outputs to improve the performance and safety of our AI models.”
Model training use of inputs/outputs is explicitly scoped to 'Trial or Research Users', implying paid or enterprise users may not have their data used for training. However, this distinction is not clearly defined elsewhere in the extracted text, creating ambiguity about which tier a given user falls into.
“This Privacy Policy does not cover the use of the Perplexity API or Enterprise Pro, where we act solely as a processor, and which are governed by the applicable Terms and Conditions, including the Data Processing Addendum incorporated therein, that you agree to when signing up fo…”
API and Enterprise Pro users have different (and potentially stronger) data-processor protections under a DPA, whereas free/Pro/Max users are governed by this policy which permits broader use of their data.