Skip to main content
CloudLast verified June 29, 2026

Perplexity data retention policy

Perplexity's Sonar API operates under a Zero Data Retention policy: prompts and responses are not retained and are not used to train models. This page covers the api.perplexity.ai path (not the consumer Perplexity app) and how to route Sonar through Meetily as a BYOK summary provider.

Quick policy snapshot

Default retention
Zero by default
Zero data retention available
Yes
Trains on API customer data by default
No

Perplexity's API training default

Perplexity's API privacy documentation is explicit for the Sonar API: it does not use customer data to train its models or for any purpose beyond processing the immediate request. The collected billing metadata (token counts, the model used, request timestamp and duration, and API key identification) does not include any content from your prompts or responses.

This is the load-bearing distinction on this page. The paid Sonar API at api.perplexity.ai is not the same as the consumer Perplexity app. For Free, Pro, and Max consumer accounts, AI training on user data is enabled by default and you must opt out in your settings. Meetily routes only through the Sonar API path under your own API key, so the consumer-app training default never applies to the transcript text Meetily sends.

Retention

Perplexity documents a Zero Data Retention policy for the Sonar API: it does not retain the data sent via the API. Prompts and responses are processed and then deleted; only billing metadata is kept, and that metadata excludes prompt and response content. Because content is not retained after the request completes, this page records the retention window for Sonar API content as zero.

The authoritative source is Perplexity's API privacy and security documentation, linked as the policy URL above. Review it before deploying any regulated workload, and note that Enterprise plans add organization-level retention controls beyond the API default.

Zero data retention

Zero Data Retention is the stated default for the Sonar API rather than an option you have to negotiate. Under it, Perplexity does not retain the data sent via the API, and that data is not used to train or improve models. For teams that need admin-managed retention windows, forced deletion, or org-wide controls, Perplexity Enterprise plans layer those settings on top, but the base Sonar API already operates under zero retention.

API vs consumer app

The privacy posture flips depending on which Perplexity product you use:

  • Sonar API (api.perplexity.ai): Zero Data Retention, no training on your content. This is the path Meetily uses.
  • Consumer Perplexity app (Free, Pro, Max): AI training on user data is enabled by default; you must opt out in Settings, and conversation retention differs from the API.

If you have read about Perplexity training on user queries, that reporting almost always concerns the consumer app, not the Sonar API. Meetily never sends your data through the consumer app, so confirm which product a given claim refers to before applying it to Meetily traffic.

How Meetily uses Perplexity

Meetily transcription runs 100% locally and audio never leaves your device. When you select Perplexity as your summary provider, Meetily uses the BYOK flow: because the Sonar API is OpenAI-compatible, you point the base URL at https://api.perplexity.ai and supply your own Perplexity API key. Meetily then sends transcript text (not audio) over TLS to the Sonar API for summarization, and the response is stored locally on your device.

Your API key means your traffic runs under Perplexity's API terms, so the Zero Data Retention and no-training posture described here applies to what Meetily sends. If you want to keep summarization on-device as well, switch your Meetily summary model to a local provider like Ollama; transcription is already local, so the entire pipeline then runs on your machine.

References

  1. "Privacy & Security," Perplexity (docs.perplexity.ai), https://docs.perplexity.ai/guides/privacy-security, accessed 2026-06-29.
  2. "Privacy & Security," Perplexity (docs.perplexity.ai legacy path), https://docs.perplexity.ai/docs/resources/privacy-security, accessed 2026-06-29.
  3. "Perplexity AI data privacy: What you need to know," Anonyome Labs (published 2026-02-05), https://anonyome.com/knowledge-center/ai-privacy/perplexity-ai-data-privacy/, accessed 2026-06-29.
  4. "Sonar API - Quickstart," Perplexity (docs.perplexity.ai), https://docs.perplexity.ai/docs/sonar/quickstart, accessed 2026-06-29.
  5. "Data Retention and Privacy for Enterprise Organizations and Users," Perplexity Help Center, https://www.perplexity.ai/help-center/en/articles/11187708-data-retention-and-privacy-for-enterprise-organizations-and-users, accessed 2026-06-29.

Last verified: June 29, 2026. Policy source: Perplexity policy

Frequently asked questions

Does Perplexity train its models on my Sonar API prompts and outputs?
No. Perplexity's API privacy documentation states that for the Sonar API it does not use customer data to train models or for any purpose beyond processing the immediate request. This page covers the paid Sonar API at api.perplexity.ai; the consumer Perplexity app has different terms (see the API-vs-consumer question below).
Is there a difference between the Perplexity Sonar API and the consumer Perplexity app for privacy?
Yes, and it is significant. The Sonar API runs under a Zero Data Retention policy with no training on your content. The consumer Perplexity app (Free, Pro, and Max accounts) enables AI training on user data by default and requires you to opt out in Settings. Meetily uses only the API path, so the consumer-app training default does not apply to traffic Meetily sends.
How long does Perplexity retain my Sonar API prompts and outputs?
Under the Sonar API Zero Data Retention policy, Perplexity does not retain the data sent via the API; prompts and responses are deleted after the request is processed. The only data retained is billing metadata such as token counts, the model used, request timestamp and duration, and API key identification - none of which includes prompt or response content. Always confirm the current terms against the linked policy.
Is zero data retention available for the Perplexity Sonar API?
Yes. Zero Data Retention is the stated default for the Sonar API, not an add-on you have to request. Perplexity documents that it does not retain data sent via the Sonar API. For organization-level controls and custom retention, Perplexity Enterprise plans offer additional admin settings.
How does Meetily handle Perplexity when I pick it as my summary provider?
Meetily transcription is always 100% local. Perplexity's Sonar API is OpenAI-compatible, so you can add it through Meetily's BYOK flow by pointing the base URL at https://api.perplexity.ai and using your own Perplexity API key. Meetily then sends transcript text (never audio) to the Sonar API for summarization. The Zero Data Retention and no-training posture on this page applies to that traffic because it is your own API key under Perplexity's API terms.

Use Perplexity with Meetily, on your terms

Meetily transcription stays 100% local. For summaries, bring your own Perplexity key (BYOK) so the data path matches the policy you just read - or pick a local model if you want zero retention by construction.

Set up BYOKDownload Meetily