Is Speechify safe? Where your documents go when it reads them aloud

The Morse teamPublished November 8, 20268 min read

As of November 2026, Speechify is safe in the ordinary sense: it's an established company that says it doesn't sell your data, encrypts stored data, and offers an optional File Encryption that locks a document on your device before it's uploaded. It isn't private by default. Documents you save are synced to its servers in the United States, cloud voices send the text to its voice servers, and its AI features send your text to OpenAI, Google, or Anthropic once you tap Allow.

We read Speechify's privacy policy (updated November 5, 2026), its terms of service (last updated July 1, 2025), its biometric data retention policy (effective October 6, 2026), its pricing page, nine of its help center articles, and its Chrome Web Store listing on November 8, 2026. We also read the permissions in the Chrome extension's own manifest, version 14.14.0. Speechify earns credit in two places. Its November 2026 policy names every outside AI company that receives your text, feature by feature, and asks before sending anything. And its File Encryption keeps a chosen file's text in a form Speechify says it can't read.

Where does your text go when Speechify reads it?

It depends on the feature. Each row comes from Speechify's own pages, read November 8, 2026.

Does Speechify store your documents?

Yes, anything you save to your Library. The privacy policy defines User Content as "all text, documents, or other content or information uploaded, entered, or otherwise transmitted by you," and says content you save "will be stored locally on that device and synced with our servers," which are in the United States (Speechify). One review that ranks for this search, dated November 6, 2022, says "Speechify doesn't record the information that it reads to you" (TechWithTech). Speechify's current policy describes the opposite for anything saved.

Speechify also says, "As a rule, Speechify employees do not monitor or view your User Content," with exceptions: a suspected terms violation, a support request you make, a legal obligation, or "to improve our algorithms as described in the User Content section of our Terms of Service" (Speechify). The terms we read have no section by that name. Their section on "User Material" covers "text, files, images, photos, video, sounds," and says "User Material is not considered to be confidential," asks you not to submit anything "in which you have any expectation of privacy," and grants Speechify an "irrevocable, perpetual" license to it, including for "promotional, or advertising purposes" (Speechify). That section's examples are profile photos and community posts, and the terms don't say whether it covers documents you only listen to. Until Speechify says, treat it as covering them.

Deleting a file moves it to a trash, where "items in the trash will be permanently deleted after 30 days" (Speechify help).

Does Speechify use your documents to train AI?

The privacy policy doesn't say it trains models on your documents, and it doesn't rule it out. It lists "to improve the features, algorithms, and usability" of its software among its uses (Speechify), and its Mac app has a "Help us Improve" option under Data and Privacy, next to a "Privacy Mode" (Speechify help). The help article doesn't say what either setting does.

For outside AI, the policy is specific. Each provider "may use your data only to provide its service to Speechify, and none may use it to train its AI models," and files you import from Google Drive aren't used to train "generalized AI/ML models" (Speechify).

Which AI companies see your text in Speechify?

Since its November 5, 2026 policy, Speechify names them. Text, documents, and images you use with Ask AI, summaries, quizzes, AI podcasts, image descriptions, and scanning go "through Speechify's servers" to OpenAI, Google (Gemini, and Cloud Vision for scanned pages), Anthropic (Claude in Ask AI), and fal for images. Voice Companion recordings go to Soniox and OpenAI, with Baseten as a backup. Text in the Speechify Keyboard's AI autocorrect goes to Inception Labs, with Anthropic as a backup (Speechify).

Speechify asks first. "Nothing is sent to the providers below until you tap Allow," and you can turn outside providers off later under Settings, then Privacy Settings. Turning it off doesn't affect "data processed before you turned it off." Users in the EEA, Switzerland, and the UK get this as a matter of consent (Speechify).

Is the Speechify Chrome extension safe?

It can read every page you open. The extension's manifest, version 14.14.0, updated November 3, 2026, asks for host access to <all_urls>, including file:// pages, and runs its scripts on all URLs, with permissions that include tabs, tabCapture, and scripting. The Chrome Web Store listing discloses that it handles "Personally identifiable information" and "Website content," and Speechify declares that the data is not sold or used for purposes unrelated to the extension's core function (Chrome Web Store).

Broad access is what a read-aloud tool needs to read any page, so this isn't a sign of misuse. It does mean the extension can see your email, bank, or health portal pages when you visit them. Chrome's extension settings let you limit its site access to the pages you choose. Our guide to whether AI browser extensions are safe covers what that permission allows.

What does Speechify do with a cloned voice?

Its biometric policy, effective October 6, 2026, covers voices created through the Speechify API, Speechify Studio, and for business customers. It collects voice samples, voice models, and verification recordings used to confirm "that the person consenting is the person whose voice is being cloned." Voice samples and models are destroyed within 30 days after the voice is deleted, consent is revoked, or the account closes, and at most 24 months after your last interaction; backup copies go within 90 days. Speechify says it doesn't "sell, lease, or trade biometric data," and keeps consent records with no biometric data for five years (Speechify). Its terms bar creating "deceptive or misleading voice clones without the explicit consent of the individual whose voice is being replicated" (Speechify). To delete a voice, email legal@speechify.com.

Is Speechify secure, and does it track you?

Speechify says it uses "industry-standard encryption to protect your data in transit and at rest" and, in its biometric policy, has "an annual independent SOC 2 Type II examination" (Speechify). If it learns of a breach, it says it "may attempt to notify you" by email or a notice on the site (Speechify). We found no breach notice on Speechify's pages as of November 8, 2026.

It doesn't sell your data or help other companies advertise to you: "Speechify does not sell or rent your Personal Data." It does advertise itself. The policy lists Meta Pixel, Google's site tag and conversion tracking, Bing Universal Event Tracking, and TikTok's conversion pixel to measure its own ads, plus Google Analytics, Hotjar, Amplitude, and Segment (Speechify). Our page on AI apps that share data with trackers explains what those pixels send.

How do you protect your documents in Speechify?

1. Turn on File Encryption before you add a sensitive file. Go to Settings, then Encryption, and set a password separate from your login. Speechify suggests it for "medical records," "contracts," and "work documents under NDA." Files already in your Library can't be encrypted later, web links can't be encrypted, and AI features and sharing are off for encrypted files (Speechify help).
2. Don't tap Allow when a feature asks to send text to an AI provider, or turn providers off under Settings, Privacy Settings (Speechify).
3. Empty the trash. Deleted files sit there for 30 days unless you delete them forever from Deleted Files (Speechify help).
4. Delete the account. Cancel any subscription first, then delete it from your profile in the app or Settings on the web (Speechify help). Speechify may keep some data for fraud, tax, and legal reasons (Speechify).

How Morse helps you ask about a document without storing it

The risks above come from your documents living in a Library synced to a company's servers, and from AI features that pass your text to other companies. Morse doesn't read text aloud. You'd use it after listening, to ask what a contract clause means, summarize a long report, or turn notes into a draft.

Attached files are sent with the message and never stored by Morse. Your chat history is stored in your browser, encrypted, not on our servers, and unlocking it takes two keys, one that never leaves your device. Morse doesn't train AI on your content, sell your data, or show ads, and runs no advertising trackers or third-party analytics. Pick a privacy mode for each chat:

• E2EE, for the medical letter or the NDA document you'd encrypt in Speechify. Paste the passage you need help with. Your browser checks the enclave's hardware attestation, then encrypts the message so only that enclave can read it, and our servers relay ciphertext they can't read. The enclaves are run by third-party providers, not by Morse, and this rests on the hardware and attestation working as designed. E2EE comes with Pro, at $19 a month.
• Private, the default, for summaries and questions about everyday reading. Open-weight models like DeepSeek, Kimi, and GLM answer here, and Morse has zero data retention agreements with its inference, search, and other providers, so they keep nothing.
• Anonymous, when you want GPT, Claude, or Gemini to explain a passage. Our proxy removes your account, email, and IP address, so the provider reads the text without knowing it's you. It may keep it for a limited time to check for abuse, under its own policy, so leave names out.

Because we don't store chats, we usually have very little to give authorities. Clearing your browser deletes your history for good and we can't recover it, so keep the original document where you already store it. Our guide to what happens to files you upload to ChatGPT covers the same question for another AI app, and is QuillBot safe covers a writing tool with a similar extension.

Speechify Premium costs $29 a month or $139 a year (Speechify). Morse's Free plan costs nothing and needs no card: start free, or try a chat from the home page without an account.