Last updated 25 October 2026. This version applies to Saywork 1.1.0 and later. What changed.
Saywork is a desktop app that types what you say and does work you ask for on your own computer. This policy explains what information Saywork handles, what it never collects, who processes it on our behalf, how long it is kept, and the rights you have. It covers the Saywork app for Windows and Linux, the Saywork Chrome helper, the Saywork Telegram bot, the Saywork connector for AI assistants, getsaywork.com and api.getsaywork.com (together, "Saywork"). Section 12 is the part about information from your Google account, written to meet Google's requirements for apps that use Google APIs.
Saywork is operated by Omar Global Ventures LLC ("we", "us"), which is the controller of the personal information described here. You can reach us at privacy@getsaywork.com.
You sign in with Google. We receive your name, email address, profile picture and Google's account identifier, and we record when you agreed to the Terms and this policy. We use them to create your account, connect your computers to it, show you who is signed in and contact you about the service. See section 12.
Your voice when you dictate, the text to be cleaned up, what you ask the agent or Ask, and what the agent reads or looks at to do the task (a web page, a file, an email you asked it to handle, a screenshot of the step it is on). This content is processed to give you the result and is not stored by us: it passes through our server to the provider that produces the answer and the answer comes back. Our server records only that a request happened, how large it was (seconds of audio, number of tokens), which model answered, how long it took and what it cost. See section 3.
Two narrow exceptions, both about security: if a request tries to make the agent reveal how it is built, we keep up to 300 characters of that request as a security alert for 90 days; and requests are logged by our hosting provider at the level of "this address called this endpoint at this time with this result", without bodies.
Your dictation history (what was heard and what was typed), notes, reminders, meetings with their transcripts and recordings, the agent's memory about you, your dictionary and snippets, Ask conversations, agent sessions, automations, settings and the app's log. These are files in your user profile on your computer. Except for the optional encrypted copy described below, we do not receive these files or back them up. The agent's detailed working record for a task, and Ask conversations, are kept encrypted with a key protected by your operating system; the working record is deleted from your computer when the task ends, while the request, the answer and the outcome remain in your history.
The Wallet. Logins, payment cards, keys and identity details you save in the Wallet are stored only on your computer, encrypted by your operating system (on Windows, DPAPI, tied to your Windows account). When you allow it, Saywork fills a value into the one website you chose for that item; the agent and the AI model see only the item's label and website, never the value, and the value is never sent to our server. The Wallet keeps a record on your computer of where and when each item was used. Resetting the Wallet in the app deletes it.
While you are signed in, the app sends events that describe what happened and never what was said: the type of event (a dictation finished, an agent task ended, a feature was used, an error occurred), language codes, modes, categories (such as "email" or "chat" for the kind of app you dictated into), durations, word counts, token counts and costs, error codes (such as "timeout"), which tool an agent step used and whether it failed, which skills were used, and the app and operating-system version. Any field that could hold your words is removed on your computer before anything is written to the send queue, and our server refuses to store an event batch that still carries such a field.
Anonymous improvement data. Separately, the app sends a second, anonymous stream a few times a day: categories and coarse counts from a closed list (for example "a dictation in Arabic, into an email app, was corrected for a misheard brand name" or "a document task failed at a step of type not-found"), with public ids of our own library rules and skills. It carries no account, device or session identifier and is sent at a random time, so it cannot be linked to you; we store only daily totals, and groups too small to be anonymous are not shown. You can turn it off in Settings → Privacy, which also deletes anything waiting to be sent.
These account events are stored with your account for 30 days so that features that depend on them — your streak, daily goal, usage allowance and the counts on your Home page — work, and are then deleted. For improving the product we use counts aggregated across people (see section 5).
For each computer: a device label, the app version and channel, the operating-system version, when it was last seen, and a one-way hash of the computer's installation identifier (see section 14). For each sign-in: a keyed one-way hash of the IP address it came from (never the address itself), the country our hosting provider reports, and the outcome. We use these to keep your account secure, to enforce one free account per computer and to detect abuse.
For each request to a speech or AI model: the time, the route (speech, cleanup or agent), the model, seconds of audio or number of tokens, whether it succeeded, how long it took and its cost. We use this to run fair-use allowances and to understand our costs.
If you connect Google, Slack, Notion or another service, we store the access grant that service gives us, encrypted, so the agent can act there when you ask. Content read from or written to those services passes through our server and is not stored or logged. If you link the Telegram bot, messages you send it wait on our server until your computer collects them, for at most 24 hours. If you create a token for an AI assistant (such as Claude) to reach your Saywork, we store only a hash of the token. See section 13.
Paid plans are sold and charged by our payment provider (section 7), which collects your payment details under its own privacy policy. We never receive your full card number. From our payment provider we receive and keep: the plan you bought, its membership and payment identifiers, its status (for example active, cancelled or past due), the renewal date, refunds, and the payment provider's identifier for you and the email address on the purchase, which we use only to match a purchase to your Saywork account.
If you invite people with your invite link or code, we keep your code, clicks on your link (the time, the country our hosting provider reports and a keyed one-way hash of the visitor's IP address, never the address), and for each person who joins: a masked hint of their email (such as "j•••@gmail.com"), the one-way computer identifiers they signed in with (section 14), a keyed hash of their sign-in address, how many days they used Saywork and whether they qualified. We use this only to count real new people fairly and to grant invitation rewards.
If you turn on "Answer me while this computer is off" (Settings → Agent; off by default), Saywork keeps an encrypted copy of your confirmed memory (excluding items marked sensitive), your notes, and the last question and answer of recent Ask chats on our server. Relevant parts are sent to our AI provider to answer you on Telegram or in a private chat where you are the only member. Shared chats cannot use your personal memory. The copy is encrypted with a key held outside the database. Server answers wait for your computer for at most 14 days, then are deleted. Turning the switch off deletes the server copy and pending answers when your computer connects. Files and folders are not included. Incognito content is not added to this copy.
The leaderboard shows your first name and last initial, activity score, streak and level to other signed-in users. You can hide your entry with the switch on the leaderboard.
Unless you choose to send it to us yourself (for example in feedback), Saywork never sends us:
Settings → Privacy in the app shows this list with what is active on your computer right now.
| When | What is sent | To | Kept by us? |
|---|---|---|---|
| You dictate, online, with cloud recognition (the default) | The recording of that dictation, and a list of your vocabulary words to spell them right | Our server → our speech-recognition provider | No. Processed and discarded; only the metering record |
| You dictate offline, or with "Keep my voice on this computer" on | Nothing (online, the recognised text may still be cleaned up in the cloud, next row) | — | — |
| Your text is cleaned up (online) | The recognised text and the style it should be written in | Our server → our language-model provider | No |
| You ask the agent, Ask, or a meeting question; a task from Telegram or an AI assistant runs | Your request, the conversation of that task, and what the agent reads or sees to do it | Our server → our language-model provider | No |
| The agent acts in a service you connected | The specific request (for example, "create this calendar event") | Our server → that service | No; the encrypted grant is kept until you disconnect |
| A task gets stuck and looks for a fix | The task, the error and a few observations, with emails, numbers and links removed | Our server → our language-model provider | Yes, 90 days, to improve the skill library |
| While you are signed in | Usage and diagnostic events, without content (section 1.4); how each agent task went | Our server | 30 days with your account; run outcomes 400 days |
| A few times a day, unless you turn it off | Anonymous improvement data: categories and counts, no identifier (section 1.4) | Our server, not linked to any account | Daily totals only, 400 days |
| Only if you turn on the speech dataset | Corrected dictations: the recording, what was heard, your correction | Our server, de-identified | Yes (section 6) |
| Only if song look-up is on and you ask what is playing | A sound fingerprint, not a recording | A song-recognition service | No |
| You send a task to the Saywork bot on Telegram | Your message and any voice note, photo or file you attach | Telegram → our server → your computer | Until your computer collects it, at most 24 hours |
| You buy or change a plan | Your payment details go to our payment provider, not to us; we receive the plan and its status | Payment provider → our server | Yes, the plan and payment status (section 9) |
| Checking for updates and downloading models | The app version | Our download server or the Microsoft Store | No |
Where the GDPR or UK GDPR applies, we rely on these legal bases:
| Purpose | Information | Legal basis |
|---|---|---|
| Provide Saywork: transcribe, clean up, run tasks, connect services, relay Telegram messages | Account; content you ask us to process; connected-service grants | Performance of our contract with you (Art. 6(1)(b)) |
| Run your account features: allowance, streaks, goals, device list | Account; usage events; metering; device information | Contract |
| Keep Saywork secure; one free account per computer; prevent abuse | Device and sign-in hashes; security alerts; metering | Legitimate interests (Art. 6(1)(f)) in protecting the service and a fair free tier |
| Fix problems and improve Saywork | Usage and diagnostic events without content, aggregated | Legitimate interests in a working, improving product; you can object (section 11) |
| Improve speech recognition with corrected dictations | The opt-in speech dataset | Your consent (Art. 6(1)(a)), which you can withdraw at any time |
| Sell and manage paid plans | Plan, membership and payment status from our payment provider; email on the purchase | Contract; legal obligation for tax and accounting records |
| Invitations and rewards | Invite code, click and join records, hashed identifiers | Legitimate interests in a fair invitation programme that cannot be gamed |
| Lifecycle and service email | Email address, name, language, usage counts | Legitimate interests (product email to existing users), with an unsubscribe link in every message; service notices under contract |
| Show your review on our website | Review, display name, photo | Your consent |
| Comply with law and respond to lawful requests | As required | Legal obligation (Art. 6(1)(c)) |
We do not make decisions that produce legal or similarly significant effects about you based solely on automated processing. Fair-use allowances are applied automatically and you can contact us about them.
Leading AI companies improve their products from how they are used. We do it without reading what you said. What we learn is of the form "dictation problem X happens in language L, and here is the fix" or "agent tasks fail at step S in apps of category A when using skills K". We get there from the content-free events described in section 1.4, counted across people, and from failure codes — never from your words, files or pages.
Speech recognition is weakest for Arabic, and the best way to improve it is examples of what was said, what the recogniser wrote and what it should have written. In Settings → Privacy you can turn on "Help improve recognition". It is off unless you turn it on, and nothing in Saywork depends on it.
We do not sell personal information, we do not "share" it for cross-context behavioural advertising (as California law defines it), and we have no advertising partners. We use service providers (subprocessors) under contract or terms to provide services to us. To limit unnecessary public exposure of our business relationships, this page describes provider categories rather than naming individual companies. The categories, purposes, information involved and processing regions are listed below. You can request current provider details at privacy@getsaywork.com. We will update this policy if a change materially affects how your information is handled or where it is processed:
| Provider category | What the service does | Information involved | Processing region |
|---|---|---|---|
| Cloud hosting and infrastructure | Hosts our server, database and storage; may maintain request logs | Information our server stores or relays | Global network; United States |
| Speech and language AI services | Speech recognition and text cleanup | Dictation audio and text, per request | United States |
| Language model services | Models behind the agent, Ask and skill research | Requests and what the agent reads, per request | United States, China and other countries |
| Account and connected-service platforms | Sign-in and services you choose to connect | Account information and your connected-service requests | United States |
| Payment and subscription services | Checkout, payments and subscriptions for paid plans | Payment details collected by the provider; plan, status and purchase email | United States |
| Email services | Sending and receiving our email | Email address, name and message | United States |
| Messaging platforms | The Saywork bot, if you link it | Messages you send the bot | Under the platform's terms |
| Website hosting | Hosts getsaywork.com | Website request logs | United States |
| Software distribution platforms | Distribution and updates through app stores | Information handled under the store's terms | Global |
Services you connect yourself (for example Slack, Notion, Linear, Dropbox, GitHub or an AI assistant) receive the requests the agent makes there on your behalf; they act under their own privacy policies. If you turn on song look-up, a song-recognition service receives a sound fingerprint.
We may also disclose information when the law requires it, to protect people's safety, rights or our service, or as part of a merger or sale of our business, in which case this policy continues to apply and we will tell you. We review any legal request and disclose only what is required. Because your content is not stored on our servers, we usually have none to disclose.
We are based in the United States and our providers process information in the countries listed in section 7, including the United States and, for some AI model requests, China. Where the GDPR or UK GDPR applies, we transfer personal information outside the European Economic Area or the United Kingdom using the European Commission's Standard Contractual Clauses (and the UK Addendum), or another lawful mechanism, together with the safeguards in section 19. Where Egypt's Personal Data Protection Law applies, we transfer data abroad only as that law permits.
Our server deletes on a schedule that runs every hour. The numbers below are the ones the code enforces.
| Information | Kept |
|---|---|
| Content processed for a result (voice, text, requests, what the agent reads) | Not stored |
| Usage and diagnostic events | 30 days |
| Detailed agent task records sent by versions before 0.9 | 30 days |
| Voice recordings sent by versions before 0.9 | 30 days |
| Corrected recordings sent by versions before 0.9 | 30 days |
| Request text on agent task records from versions before 0.9 | Cleared after 30 days |
| How each agent task went (outcome, steps, timings, failure codes) | 400 days |
| Usage metering | 400 days |
| Screenshots attached to feedback | 180 days |
| Stuck-task research records (scrubbed) | 90 days |
| Proposed skill fixes that were not adopted | 90 days |
| Security alerts | 90 days |
| Link between a speech-dataset item and your account | 30 days |
| De-identified speech-dataset items | 2 years |
| Telegram messages waiting for your computer | Until collected, at most 24 hours |
| Sign-in address hashes | 90 days |
| Website visit counts | 35 days (daily totals kept) |
| Anonymous improvement totals (not linked to anyone) | 400 days |
| Account, feedback, writing voice, connected-service grants, reviews | Until you delete them or your account |
| Plan, membership and payment status | While your plan is active, then as long as tax and accounting law requires; kept after account deletion only for that purpose |
| Google access grant not used for 180 days | Revoked at Google and deleted |
| Clicks on an invite link (hashed address, country) | 90 days |
| Who joined with an invite, and whether they qualified | Kept after account deletion, under a Google account identifier rather than your email, so the same person is never counted twice |
| After you delete your account | Everything above that is linked to you is deleted; storage is cleared within the hour. Two things remain: the hashed computer identifier bound to a hash of your Google account (so a reinstall cannot reset the free allowance), and your last 24 hours of usage counts for 48 hours. |
Depending on where you live, you have some or all of these rights over personal information we hold: to know and access it; to receive it in a portable format; to correct it; to delete it; to restrict or object to processing (including processing based on legitimate interests); to withdraw consent at any time without affecting earlier processing; and not to be discriminated against for using them. Export everything and Delete everything in the app fulfil access, portability and deletion immediately. You can also write to privacy@getsaywork.com; we will verify the request comes from the account holder (normally by replying from the account's email address) and answer within one month, or sooner where the law requires.
The rights in Articles 15–22 of the GDPR and the UK GDPR apply. You may lodge a complaint with your local data protection authority or, in the UK, the Information Commissioner's Office. We would appreciate the chance to address your concern first.
Under the CCPA as amended by the CPRA, California residents have the rights to know, delete, correct, and opt out of sale or sharing, and to limit use of sensitive personal information. In the last 12 months we collected the categories identifiers (name, email, device identifiers), internet or network activity (usage events), audio information (only as processed for a result, or the opt-in dataset) and inferences (your writing-voice profile), from you and your devices, for the purposes in section 4. We do not sell or share personal information, and we use sensitive personal information only to provide the service you request. You may use an authorised agent; we may ask them for proof of authority. Residents of other states with comparable laws (including Colorado, Connecticut, Virginia, Texas, Oregon and Montana) have similar rights and may appeal a refusal by replying to our answer; if we deny the appeal you may contact your state attorney general.
Under Personal Data Protection Law No. 151 of 2020, you have the right to know about, access and obtain your personal data held by any controller; to withdraw prior consent; to correct, amend, erase, add to or update it; to restrict processing to a certain scope; to be notified of any breach of your personal data; and to object to processing that conflicts with your fundamental rights and freedoms. You may complain to the Personal Data Protection Center.
This section describes how Saywork accesses, uses, stores, shares, keeps and deletes information it receives from your Google account through Google APIs ("Google user data"). It applies on top of the rest of this policy; where they differ, this section wins for Google user data.
Signing in asks Google only for your name, email address, profile picture and account identifier (the openid, email and profile permissions). Nothing else from your Google account is accessed unless you connect a Google product yourself, on the Connections page, and approve it on Google's own consent screen. Each product is connected separately, and asks only for the permissions in its row:
| Product | Permission (scope) | What Saywork does with it | What it never does |
|---|---|---|---|
| Sign-in | openid, email, profile | Creates your account and shows who is signed in | Reads nothing else in your Google account |
| Google Calendar | calendar.events | Reads your events to answer you; creates, changes or deletes an event only after you approve that exact change; names a meeting you are in, if you turn that on | Change calendar settings or sharing |
| Gmail | gmail.readonly, gmail.compose | Searches and reads mail when you ask; writes drafts; sends a message only after you approve the recipients and text | Delete, label or archive mail, or change your settings |
| Google Docs | documents | Reads a document you name; creates one or adds to one when you ask | Open documents you did not name |
| Google Sheets | spreadsheets | Reads a sheet you name; creates one, adds or updates rows when you ask | Open sheets you did not name |
| Google Drive | drive.file | Lists and reads only files Saywork created or you opened with it; shares one only after you approve | See the rest of your Drive |
| Google Meet | meetings.space.created, meetings.space.readonly | Creates a meeting link when you ask; reads a past meeting's record, participants and, where your Google plan provides one, its transcript, to write your notes | Join or record a meeting through Google |
Google user data is used only to provide the features you use and that are described in the table above, at the moment you ask for them: to answer your question, or to do the task you asked for. Anything that another person would see — sending an email, sharing a file, inviting someone to an event — is shown to you first and happens only after you approve it.
Saywork's agent is powered by an AI language model. When a task you ask for needs the content of Google user data — for example, to summarise an email you asked about, or to write a reply to it — that content is sent, together with your request, through our server to a third-party AI model provider (section 7) solely to produce that result, and the result comes back to you. It is not stored by us, and the provider processes it only to return that result, under its API terms. We do not use Google user data to develop, improve or train generalised AI or machine-learning models, ours or anyone else's.
We do not sell Google user data, use it for advertising, or use it to decide anyone's creditworthiness or eligibility for anything. We transfer it only: to provide the features above (to Google itself, and to our AI model provider as described in 12.3); to comply with law; to protect against fraud, abuse or security threats; or as part of a merger or sale of our business, with notice to you. Our cloud hosting provider relays it as part of running our server.
No person at Saywork reads your Google user data, unless you ask us to look at a specific item and agree to it, it is needed to investigate abuse or a security incident, or the law requires it.
Disconnect on the Connections page at any time: we revoke our access at Google and delete the stored grant immediately. Deleting your account does the same for every Google product. You can also remove Saywork at myaccount.google.com/permissions. Because we do not store content from your Google account, there is nothing else for us to delete; what you saved on your own computer stays yours, and you can delete it there.
Separately from the connections above, the agent can work in Google websites — Gmail, Calendar, Docs and others — that are open in your own Chrome, where you are already signed in, the same way you would, using the Saywork Chrome helper. That does not use Google APIs or any Google permission given to Saywork. What the agent reads there to do your task is handled like any other page it reads: sent to our AI model provider to produce the result, and not stored by us (section 3).
Saywork's use and transfer to any other app of information received from Google APIs will adhere to the Google API Services User Data Policy, including the Limited Use requirements. For Google Workspace APIs, Saywork also follows the Google Workspace API User Data and Developer Policy.
Services you connect (Slack, Notion, Linear, Trello, Todoist, Dropbox, HubSpot, Figma, ClickUp, Zoom, X and others): connected only by you, on the service's own consent screen, with only the permissions the listed features use. The grant is stored encrypted on our server and used only to do what you ask; anything that sends, posts, shares, changes or cancels something asks you first. Content passes through and is not stored. GitHub is connected from the app and its key stays on your computer, encrypted by your operating system. Apps you have not connected (such as WhatsApp) are used by the agent in your own app or browser on your computer; we hold nothing for them.
Telegram. If you link the Saywork bot, the messages you send it (text, and voice notes, images or files you send for a task) are received by our server and held until your computer collects them, at most 24 hours, then deleted. Your computer does the work and replies through the bot; replies pass through and are not stored.
AI assistants (MCP). If you create a token so an AI assistant such as Claude can reach your Saywork, we store only a hash of the token. When the assistant asks for something (for example, to search your notes), the request is relayed to your computer, your computer answers, and the answer is handed back and deleted from our server as soon as it is delivered. The assistant's provider handles what it receives under its own policy.
Saywork's free allowance is per person. To reduce resets through new accounts or reinstalls, the app sends a cryptographic hash of the computer's installation identifier (on Windows, the machine GUID). We bind it to the account's stable identity. Verified email mappings let supported sign-in methods use the same account and allowance. After account deletion we remove content and password credentials but retain minimal account identity, verified email mappings, computer bindings and carried usage counts to prevent allowance resets. These records can link sign-in methods; they are not anonymous. Sign-in IP addresses are kept only as keyed hashes for 90 days. You can ask us to release a computer you sold or gave away.
We send service email (for example about your account or security) and lifecycle email (a welcome, tips in your first week, a weekly summary of your usage counts, and a note if you have not used Saywork for a while), in your app's language, through our email provider. Lifecycle email uses your name, email, language, time zone and counts such as your streak or time saved — never your content. Every lifecycle email has an unsubscribe link that works without signing in. When you write to hello@ or privacy@getsaywork.com, your message and our reply are kept in our mailbox; our email provider reports delivery and, where your mail client allows it, opens and clicks of email we send, which we use only to know a message arrived.
If you write a review in the app, you choose the display name and whether to add a photo (uploaded, or your Google picture), and you tick a separate box to allow us to publish it on getsaywork.com. We publish only reviews you agreed to publish and we approved. You can withdraw or delete your review at any time, and deleting your account deletes it and its photo. Feedback you send is used to fix Saywork and is kept with your account; screenshots you attach are deleted after 180 days.
getsaywork.com loads the Whop Pixel to measure page visits and checkout attribution. Loading this third-party script sends browser request information, including your IP address and browser information, to Whop. App memory, notes, dictations and Google integration content are not sent to this pixel. See Whop’s privacy policy. Separately, our own analytics count page views and download clicks: each count carries the page, its language, the host name of the referring site and any campaign tag. To count a visitor once a day, the IP address and browser are combined with a random value replaced every day and deleted after two days, and only the resulting hash is stored. Browsers sending Do Not Track or Global Privacy Control are not counted. Counts are deleted after 35 days; daily totals are kept. Fonts are loaded from Google Fonts, which receives your IP address when you load a page. The download section of the English and Arabic home pages also loads the official Microsoft Store badge: its script, images and an interactive frame come from Microsoft's servers at get.microsoft.com, which receive the same browser request information, including your IP address; opening the badge takes you to Microsoft, and what Microsoft does there is subject to Microsoft's privacy statement, not this one.
Saywork is for adults. It is not directed to anyone under 18 and we do not knowingly collect personal information from anyone under 18, or from children under 13 in any circumstances. If you believe a person under 18 is using Saywork, write to privacy@getsaywork.com and we will delete the account and its information.
Everything between the app, our server and our providers travels over TLS. Connected-service grants are encrypted with a key held outside the database. On your computer, the agent's working files, Ask conversations and queued task records are encrypted with a per-install key protected by your operating system, and the Wallet is encrypted by your operating system. The admin panel has its own sign-in, separate from user accounts, with an authenticator code as a second factor. Access to production systems is limited to the people who run Saywork and is logged. We remove content fields on your computer before anything is queued, and our server re-checks every upload against an allow-list. Our practices and the standards we align with are on the Trust & security page. No system is perfectly secure; if a breach affects your personal information we will notify you and the relevant authorities as the law requires. Report a vulnerability to security@getsaywork.com.
When you record a meeting or dictate near others, their voices may be captured. Meeting recordings and transcripts stay on your computer; when you transcribe a meeting with cloud recognition, its audio is processed like dictation and not kept. Recording conversations can require the other people's knowledge or consent depending on where you and they are; please make sure you have it.
25 October 2026: clarified how verified sign-in methods share an account and which minimal identity, email, device and usage records remain after account deletion. Password credentials and account content are deleted.
This version (11 October 2026) replaces the policy of 3 October 2026. It adds: the Wallet, which stays encrypted on your computer (section 1.3); paid plans and our payment provider (sections 1.9, 7 and 9); invitations (section 1.10); Telegram tasks in the list of what leaves your computer (section 3); and a fuller account of how we handle information from Google, including the use of an AI model to produce what you ask for (section 12). Nothing that was previously not collected is now collected from your content.
The version of 3 October 2026 replaced the policy of 26 September 2026. The main changes: Saywork no longer collects dictation recordings, transcripts or agent requests; using Saywork no longer requires agreeing to send voice recordings; sharing corrected dictations is now a separate opt-in; usage events carry no content and are deleted after 30 days; and we added retention periods, provider categories, your rights by region, and the Trust & security page. Recordings, transcripts and task records collected under the previous policy are deleted on the schedule in section 9. We will tell you in the app or by email before a change that reduces your rights or expands what we collect, and we will ask for your consent where the law requires.
Omar Global Ventures LLC (Saywork) — privacy@getsaywork.com. Security reports: security@getsaywork.com.