Privacy Policy

Version 2.3 Effective Date: August 26, 2026

Introduction

This Privacy Policy describes how Avlok AI, LLC (“Avlok,” “we,” “us,” or “our”) collects, uses, discloses, and protects personal information in connection with the Avlok application and related services (collectively, the “Service”). It also explains the choices and rights available to you.

Avlok is a photo and video search application that lets you find media in your personal library by describing it in ordinary language. We designed Avlok so that privacy is enforced by the architecture of the product rather than by promise alone. The central principle is simple: your original, full-resolution photos and videos never leave your device. The cloud components of the Service operate on small, metadata-stripped derivatives, and most features run entirely on your computer.

This policy is written to comply with applicable data protection laws, including the EU General Data Protection Regulation (“GDPR”), the UK GDPR and Data Protection Act 2018, the California Consumer Privacy Act as amended by the California Privacy Rights Act (“CCPA/CPRA”), and other United States state privacy laws. Defined terms are used as set out throughout this policy.

By creating an account or using the Service, you acknowledge that you have read and understood this policy. Where we rely on your consent, we will ask for it separately and you may withdraw it at any time as described below.

1. Who we are and how to contact us

The data controller (for GDPR and UK GDPR purposes) and the business (for CCPA/CPRA purposes) responsible for your personal information is:

Avlok AI, LLC 502 West 7th St, STE 100 Erie, PA 16502-1333 United States

For any privacy question, request, or concern, contact us at privacy@avlok.ai. For general inquiries, contact support@avlok.ai. We aim to acknowledge privacy requests within five (5) business days and to fulfill them within the periods required by applicable law.

EU and UK representatives. Individuals in the European Economic Area (“EEA”) and the United Kingdom may contact us directly at the address above or at privacy@avlok.ai.

Data Protection Officer. We are not currently required to appoint a Data Protection Officer. If that changes, the DPO’s contact details will be published here. In the meantime, privacy matters may be directed to privacy@avlok.ai.

2. Scope of this policy

This policy applies to personal information we process when you create an Avlok account and use the Service. It does not apply to third-party products, websites, or services that we do not control, even where they are referenced in or accessible through the Service. Those third parties maintain their own privacy practices, and we encourage you to review them.

3. The information we collect

We collect only the information needed to operate the Service. The categories below describe what we collect, where it is held, and whether it ever leaves your device.

3.1 Information you provide when you register

Email address. Used to identify your account, authenticate you, and send you operational messages about the Service.
Password. Stored only as a one-way cryptographic hash (bcrypt). We never store, and cannot recover, your plaintext password.
First name and last name. Used to personalize account communications (for example, the greeting in a welcome email). Stored on your account record and shared only with our email-delivery provider to the extent needed to send those messages.
Marketing preference. A single indicator of whether you have opted in to receive product announcements. This defaults to off. We record the date and time of an opt-in as an audit record of consent.
Referral source (optional). A single selection (or a short free-text response) describing how you learned about Avlok. Used only for product analytics. You may decline to provide this.

3.2 Information we derive at registration without storing the source

Country of registration. At the moment you create an account, our servers briefly read the IP address of your request and resolve it to a two-letter country code (for example, US, DE, JP) using a copy of MaxMind’s GeoLite2-Country database that is held within our own infrastructure. We store only the resulting country code, which we use to determine which data protection laws apply to your account. Your IP address is not stored in our database and is not sent to MaxMind or any other third party for this lookup.
Age confirmation. At registration we ask for your date of birth solely to confirm that you meet the minimum age described in Section 17. We validate it during the request and discard it; we retain only a flag recording that age was confirmed.

3.3 Information generated from the photos and videos you index

When you ask Avlok to index a folder or connect your Apple Photos library, the Service processes your media to make it searchable.

Apple Photos library (optional). If you choose to connect your Apple Photos library, Avlok reads reduced-size previews and item information (such as capture dates, locations, and dimensions) from the Photos app on your device, using the photo-library permission that macOS asks you to grant. This produces the same information described in this section as indexing a folder, and nothing more. Avlok never changes, moves, or deletes anything in your Photos library. If you disconnect it, everything Avlok imported from it is removed from Avlok. The following information is generated:

Small thumbnails. Avlok creates a 224 by 224 pixel JPEG thumbnail of each photo (and of representative frames of each video). All embedded metadata, including any EXIF data such as GPS coordinates and camera information, is removed from the thumbnail on your device before it is transmitted.
Search representations (“embeddings”). A numeric representation (a list of floating-point values) is generated from each thumbnail to enable description-based search. This representation is mathematically derived and cannot be used to reconstruct the original image. It is stored on your device and, if you enable Cloud AI Access, on our servers.
Content descriptions. If you enable an AI feature that uses them (the local AI integration or Cloud AI Access), Avlok generates a short text description of the visible content of each photo from its metadata-stripped thumbnail, using the cloud processing providers described in Sections 7 and 8. The description is stored on your device. A copy of the text of each description is also retained on our servers, associated with your account and the file fingerprint of the photo, so that we can return it without reprocessing the photo if the same photo needs a description again (for example, after a reinstall or on a second device). This copy contains only text and a fingerprint; it contains no image data, and it is deleted when you delete your account.
File fingerprints. A cryptographic hash (SHA-256) of each source file, used to detect changes and avoid reprocessing unchanged files.

3.4 Information held only on your device (never transmitted to us)

The following information is created and used entirely on your computer. It is stored in your local search index, it is not transmitted to Avlok, and we have no access to it:

Your original, full-resolution photos and videos, and their original metadata in unredacted form.
People grouping (face data). If you use Avlok’s people-grouping feature, the Service detects faces in your photos and generates on-device representations (which may include face geometry) used to group photos of the same person. These face representations and any face geometry are stored only on your device, are never transmitted to Avlok, and are not accessible to us. However, if you separately enable Cloud AI Access, a limited set of derived people information (the names you assign to people, and which of your synced photos each named person appears in) is synced to our servers so that connected AI assistants can find photos of a named person. That derived information is described in Sections 3.5 and 7; the underlying face representations and face geometry still never leave your device. Section 6 contains additional information about this feature.
Spoken-words video transcripts (audio and transcription). If you enable spoken-words video search (a Pro feature), Avlok reads the audio track of your videos on your device and generates short text transcripts of the spoken content. The audio itself is never uploaded, and the transcription process runs entirely on your device. The resulting transcripts are stored in your local search index. If you separately enable Cloud AI Access, the text of the transcripts for the videos you choose to sync is stored on our servers, as described in Sections 3.5 and 7, so that connected AI assistants can search what was said in your videos and so that your other devices can search those spoken words without re-processing the video. If you do not enable Cloud AI Access, transcripts remain on your device only.

3.5 Information generated only if you enable Cloud AI Access

Cloud AI Access is an optional Pro feature that stores a copy of your search index in our cloud so that connected AI assistants can search your library from any device. If, and only if, you enable it, the following are stored on our servers for the photos you choose to sync:

The search representations (embeddings) described in Section 3.3.
Short text descriptions of the visible content of each photo, generated by our cloud processing.
A 224 by 224 pixel, metadata-stripped thumbnail of each synced photo.
A defined set of metadata fields associated with each synced photo: the date the photo was taken, the camera and lens information, the image dimensions, a coarse location name (such as a city and country), and the underlying GPS coordinates, where present. The GPS coordinates are stored to power location-based search within the Service.
Spoken-words transcripts, only if you also use spoken-words video search. If you have enabled both Cloud AI Access and spoken-words video search (Sections 3.4 and 6), the text transcripts of the spoken content of your synced videos, together with the timestamps of each spoken segment, are stored on our servers. We do not receive or store the audio itself: only the text your device produced. This is what allows a connected AI assistant to answer a request such as “find the video where she mentions the deadline,” and what allows your other devices to search those spoken words without re-processing the video.
Derived people information, only if you also use people grouping. If you have enabled both Cloud AI Access and the people-grouping feature (Section 6), the following are synced: the names you assign to people, and, for each named person, which of your synced photos they appear in, together with the co-appearance groupings used to answer questions such as “who appears with whom.” We do not sync the underlying face representations, face geometry, or any face images: only the names you chose and the groupings they label. This derived information is what allows a connected AI assistant to answer a request such as “show me photos of Sarah.” Because it is derived from biometric processing, we treat it as sensitive/special-category information; see Sections 6, 7, and 12.

We describe the cloud data flows in detail in Section 7.

3.6 Search queries

When you search your library, Avlok interprets your query so it can return relevant results. Many queries are interpreted entirely on your device. For certain queries, the Service sends the text of your query, together with selected labels from your library used to interpret it (for example, names you have assigned to people, places, or trips), to our cloud query-interpretation provider. The result is a structured understanding of your query, which is used to retrieve matching media. We do not use your queries to build an advertising profile, and we do not sell them.

3.6a Assistant search history (only if you use a connected AI assistant with Cloud AI Access)

When you run a search through a connected AI assistant that has Cloud AI Access enabled (for example, Claude on the web, Claude on your phone, or ChatGPT), Avlok temporarily stores two pieces of information so that you can re-open the same result set in the Avlok app on your computer: the text of the search you ran through the assistant, and a list of cryptographic hashes that point to the specific photos in the result set on your device. These records do not contain your images, do not contain any image data, and are limited to your most recent twenty searches.

Each record is automatically deleted after seven days. Records are also deleted immediately if you delete your cloud data or close your account.

This feature is part of Cloud AI Access. If you do not enable Cloud AI Access, no assistant-search records are created. If you enable Cloud AI Access and later disable it, your assistant-search records are deleted at the same time the rest of your cloud data is deleted, as described in Section 12.

3.7 Usage and billing information

Usage counts. The number of photos indexed in a billing period, used to enforce plan limits.
Subscription information. Your plan, subscription status, and billing period dates, together with identifiers issued by our payment processor (a customer identifier and a subscription identifier). Payment card details are handled entirely by our payment processor; Avlok never receives or stores your card number.

3.8 Technical and security information

IP address (transient). Your IP address is processed at our network edge and backend to deliver the Service securely, to apply rate limits on sign-in and account-creation requests, and (at registration only) to resolve your country code as described in Section 3.2. We do not store your IP address in our database and we do not include it in our analytics.
Authentication and authorization tokens. Tokens used to keep you signed in and to authorize connected AI assistants. Where these are stored, they are stored as cryptographic hashes rather than in plaintext.

3.9 Anonymous product analytics

If you do not opt out, Avlok collects anonymous, non-identifying usage events to help us understand how the Service is used and to improve it. Each event consists of an event name (for example, “search executed”), a small set of bucketed properties that describe the shape of an action rather than its content (for example, that a query contained three words, never the words themselves), and a random identifier created on first launch that is not linked to your account, your email, or any other information we hold about you. The technical safeguards that prevent personal information from reaching our analytics provider are described in Section 15.

4. Information we do not collect, and our core commitments

This section is as important as the one above. We commit that:

We do not collect or store your original, full-resolution photos or videos. Only metadata-stripped 224 by 224 pixel thumbnails are transmitted for processing, and those are discarded immediately after the search representation is generated (see Section 7).
We do not receive your video audio. Spoken-words transcription runs entirely on your device. If you enable Cloud AI Access, the text transcripts your device produced for synced videos are stored on our servers as described in Sections 3.5 and 7, but the audio itself never leaves your device.
We do not receive your face geometry or the underlying on-device face representations; face detection and grouping run entirely on your device. If you enable Cloud AI Access, a limited set of derived people information (the names you assign and which photos each named person appears in) is synced so that connected AI assistants can search by person, as described in Sections 3.5 and 7, but the face representations and geometry themselves still never leave your device.
We do not store your date of birth. We confirm your age at registration and then discard it.
We do not use your photos, videos, captions, transcripts, search representations, or any of your content to train artificial-intelligence or machine-learning models.
We do not sell your personal information, and we do not share it for cross-context behavioral advertising.
We do not disclose your personal information to advertisers.
We do not include personal information in our anonymous analytics.

For users who do not enable Cloud AI Access, the metadata associated with your photos (including GPS coordinates, capture dates, and camera information) remains on your device and is not transmitted to us. If you enable Cloud AI Access, a defined set of those metadata fields is synced to our servers as described in Sections 3.5 and 7.

5. How and why we use your information, and our legal bases

We use personal information for the purposes set out below. For individuals in the EEA and the United Kingdom, the table identifies the legal basis on which we rely under Article 6 of the GDPR.

Purpose	Information used	Legal basis (EEA / UK)
Create and manage your account; authenticate you; provide local and Pro search features you have requested	Account information; passwords; usage counts	Performance of a contract (Art. 6(1)(b))
Generate search representations from thumbnails, and interpret your search queries, in order to deliver search results	Thumbnails; search representations; query text and interpretation labels	Performance of a contract (Art. 6(1)(b))
Provide Cloud AI Access (storing your cloud search index and serving it to connected AI assistants)	Cloud-synced data described in Section 3.5	Your consent, given by enabling the feature (Art. 6(1)(a)); withdrawable at any time
Process payments and manage subscriptions	Subscription information; payment-processor identifiers	Performance of a contract (Art. 6(1)(b))
Send transactional email (for example, welcome, password reset, deletion confirmation)	Email address; first name	Performance of a contract (Art. 6(1)(b))
Send marketing communications	Email address; marketing preference	Your consent (Art. 6(1)(a)); withdrawable at any time
Confirm your age and determine which laws apply to your account	Date of birth (discarded); country code	Compliance with a legal obligation (Art. 6(1)(c)); our legitimate interests in lawful operation (Art. 6(1)(f))
Secure the Service, prevent fraud and abuse, and apply rate limits	IP address (transient); authentication tokens	Our legitimate interests in protecting the Service and our users (Art. 6(1)(f))
Understand and improve the Service through anonymous analytics	Anonymous events; random identifier	Your consent (Art. 6(1)(a)); withdrawable at any time

Where we rely on legitimate interests, we have balanced those interests against your rights and freedoms, and we will provide further information about that balancing on request.

Providing your account information is necessary to create an account and use the Service; if you choose not to provide it, we will not be able to provide the Service to you. Optional features (Cloud AI Access, marketing communications, analytics, and the referral source) are not necessary to use the core Service.

6. Processing that happens only on your device

A defining characteristic of Avlok is that most processing occurs on your own computer. The following features run entirely on your device, and the data they produce is stored only in your local search index unless this policy states otherwise:

Local search. Description-based search of your library runs against the local search index on your device.
People grouping. This feature is off by default and runs only if you choose to turn it on. When enabled, Avlok detects faces in your photos and creates on-device representations (which may include face geometry) to group photos of the same person, together with any names you assign. The face representations and any face geometry are created, stored, and used only on your device; they are never transmitted to Avlok, and we cannot access them. Because face geometry may be treated as biometric information under certain laws (including the GDPR’s special-category rules and various United States state biometric statutes, such as the Illinois Biometric Information Privacy Act), we note expressly that Avlok does not collect, capture, receive, store, or otherwise obtain the face representations or geometry; they remain under your control on your device. The sole purpose of this feature is to let you find and organize photos of specific people within your own library. We do not use it to identify strangers, we do not use it for advertising, we do not sell or profit from it, and we do not use it to train any model. If you separately enable Cloud AI Access, the names you assign and the groupings they label (never the face representations or geometry) are synced so that connected AI assistants can find photos of a named person, as described in Sections 3.5 and 7. You may rename or clear names, hide or delete a group, disable the feature, or delete your account at any time; disabling the feature or deleting your account or your cloud data removes any synced people information within the periods described in Section 10. If you do not wish to use this feature, you may leave it disabled.
Spoken-words video transcription. When enabled, Avlok transcribes the spoken content of your videos on your device. The audio is never uploaded. The transcripts live in your local search index; if you separately enable Cloud AI Access, the transcript text for synced videos is additionally stored on our servers as described in Sections 3.5 and 7.

These on-device features are not affected by whether you enable Cloud AI Access; they remain on your device in all cases.

7. When information leaves your device

The Service uses a limited number of cloud components. The flows below are the only circumstances in which information derived from your library is transmitted from your device.

a. Generating search representations and content descriptions. When you index a folder or connect your Apple Photos library, Avlok creates a small, metadata-stripped thumbnail of each item on your device and transmits it over an encrypted connection to our cloud processing providers, which generate the search representation and, where an AI feature that uses them is enabled, the short content description, and return them to your device. The thumbnail is processed transiently and discarded at the end of the request; it is not stored by any processing provider. The text of content descriptions is retained as described in Sections 3.3 and 10.

b. Cloud AI Access (optional). If you enable Cloud AI Access, the data described in Section 3.5 (search representations, content descriptions, metadata-stripped thumbnails, a defined set of metadata fields including coarse location and GPS coordinates, and, where the respective features are enabled, spoken-words transcript text and derived people information) is stored on our servers so that connected AI assistants can search your library across devices. You may disable Cloud AI Access at any time; the consequences of doing so are described in Section 10.

c. Interpreting search queries. For certain queries, the text of your query and selected interpretation labels from your library are transmitted to our cloud query-interpretation provider, which returns a structured understanding of the query. This enables Avlok to deliver relevant results for queries that cannot be fully interpreted on your device.

d. Delivering results to a connected AI assistant (optional). If you enable Cloud AI Access and connect a third-party AI assistant (for example, ChatGPT, provided by OpenAI, or Claude, provided by Anthropic) and then use that assistant to search your library, the results of your searches are transmitted to that assistant’s provider so the assistant can present them to you. Depending on what you ask, those results may include: metadata-stripped thumbnails of the matching photos; the short content descriptions Avlok generated for them; capture dates, camera information, and image dimensions; location information, including city, country, and GPS coordinates; the names you have assigned to people, places, and trips, and the groupings they label; excerpts of the spoken-words transcripts of matching videos, with their timestamps, where you use spoken-words video search; and the text of the request you made to the assistant. You choose whether to connect an assistant and which one. The provider of the assistant you connect processes this information under its own privacy policy and terms, which are separate from this policy and outside our control; we encourage you to review them. As above, your original, full-resolution photos and videos are never transmitted; only metadata-stripped thumbnails are shared for display. If you do not connect an AI assistant, none of this applies.

In each case, your original, full-resolution photos and videos remain on your device and are never transmitted.

8. Sub-processors and other recipients

We use the service providers below (“sub-processors”) to operate the Service. Each processes personal information only on our instructions and under contractual obligations consistent with applicable law.

Provider	Role	What it receives
Modal (modal.com)	Cloud processing (generating search representations; fallback provider for content descriptions)	Metadata-stripped thumbnails, processed in memory and discarded after the request
Supabase (supabase.com)	Database, authentication, and file storage	Account information; usage; authentication and authorization records; the content-description text described in Section 3.3; and, for users who enable Cloud AI Access, the cloud-synced data described in Section 3.5
Amazon Web Services (aws.amazon.com), United States region	Cloud query interpretation; primary provider for content descriptions	The text of certain search queries and selected interpretation labels (such as names you assign to people, places, or trips); metadata-stripped thumbnails, processed transiently to generate content descriptions and not stored
Cloudflare (cloudflare.com)	Network edge and security (reverse proxy)	Requests to our servers pass through Cloudflare’s edge, where encryption is terminated; Cloudflare therefore transmits request content (including thumbnails, sign-in details, and query text) but does not retain request bodies in the ordinary course
Stripe (stripe.com)	Payment processing	Email address and payment details collected directly by Stripe; Avlok never receives your card number
Resend (resend.com)	Transactional and (where you opt in) marketing email delivery	Email address and, where present, first name
PostHog (posthog.com), United States	Anonymous product analytics	Anonymous event names, bucketed properties, and a random identifier not linked to your account; no email, query text, file paths, or stored IP address
Have I Been Pwned (haveibeenpwned.com)	Breached-password check at registration and password change	A 5-character prefix of the SHA-1 hash of your password (not the password itself, and not enough on its own to identify any particular password). The rest of the hash is matched locally against the returned list of known-breached hashes; no further data is transmitted. The service is operated by an independent Australian provider and fronted by Cloudflare’s network.

Each provider maintains its own privacy practices: Modal, Supabase, Amazon Web Services, Cloudflare, Stripe, Resend, PostHog, Have I Been Pwned.

We may engage additional or replacement sub-processors as the Service evolves. We maintain an up-to-date record of our sub-processors and will update this policy when our use of sub-processors materially changes. We may also disclose personal information where required by law, to enforce our agreements, to protect the rights, safety, and property of Avlok or others, or in connection with a merger, acquisition, financing, or sale of assets, in which case we will require the recipient to honor this policy or provide you with notice.

Connected AI assistants (recipients you choose). Separately from the sub-processors listed above, if you enable Cloud AI Access and connect a third-party AI assistant, the provider of that assistant, such as OpenAI (for ChatGPT) or Anthropic (for Claude), receives the search results described in Section 7(d) when you use it to search your library. These providers are not our sub-processors: they act on your instruction as the user who connected them, and they process the information under their own privacy policies and terms, which we do not control. You may disconnect an assistant, or disable Cloud AI Access, at any time.

9. International data transfers

Avlok is based in the United States, and our sub-processors process personal information in the United States and potentially in other countries. If you are located in the EEA, the United Kingdom, or another jurisdiction with cross-border transfer restrictions, your personal information will be transferred to, and processed in, countries that may not provide the same level of data protection as your home jurisdiction.

Where we transfer personal information out of the EEA or the United Kingdom, we rely on a lawful transfer mechanism, which may include the European Commission’s Standard Contractual Clauses (and the UK International Data Transfer Addendum), an adequacy decision, or, where applicable, a provider’s certification under the EU-U.S. Data Privacy Framework and its UK extension. You may request more information about the safeguards we apply by contacting privacy@avlok.ai.

10. Data retention

We keep personal information only for as long as needed for the purposes described in this policy, or as required by law.

Account information is retained until you delete your account.
Cloud AI Access data is retained while the feature is enabled. If you disable it, the data is retained for a grace period of fourteen (14) days so that you can re-enable the feature without rebuilding your cloud index, after which it is permanently deleted by an automated process. A daily automated process also removes cloud data associated with deleted accounts as a backstop.
Synced people information (the names and groupings described in Section 3.5, present only if you use people grouping together with Cloud AI Access) is retained on our servers only while Cloud AI Access is enabled. It is subject to the same fourteen (14) day grace period and permanent deletion on disablement as other Cloud AI Access data, and it is deleted when you delete your account. The underlying face representations and face geometry are never stored on our servers; on your device, they are retained until you clear the names, delete the group, disable the feature, or uninstall the application.
Spoken-words transcripts synced under Cloud AI Access (Section 3.5) follow the same lifecycle as other Cloud AI Access data: they are retained while the feature is enabled, subject to the same fourteen (14) day grace period on disablement, removed when you stop sharing the folder containing the video, and deleted when you delete your cloud data or your account. The audio itself is never stored on our servers. On your device, transcripts persist in your local search index until you delete the video or uninstall the application.
Assistant search history described in Section 3.6a is retained for seven (7) days from the time of the search, after which it is automatically deleted. A daily automated process removes records past this retention period. The information is also deleted immediately if you delete your cloud data or close your account, and the retention period does not extend if you sign in from a different device. Only the twenty most recent searches per account are kept; older records are removed as new searches are recorded.
Thumbnails sent for processing are discarded immediately after processing and are not retained.
Content descriptions (the text copies described in Section 3.3) are retained on our servers, associated with your account, until you delete your account, at which point they are deleted with the rest of your account data. For Cloud AI Access users, the copy stored with synced photos additionally follows the Cloud AI Access retention above.
Local data on your device persists on your device until you delete the relevant files, delete your account, or uninstall the application.
Password-reset codes expire fifteen (15) minutes after they are issued.
Payment records held by our payment processor are retained in accordance with that processor’s policies and applicable financial and tax laws; following account deletion, that processor may retain anonymized records as required by law.
Anonymous analytics are retained in accordance with our analytics provider’s retention practices and can be deleted by reference to the random identifier described in Section 15.
Backups. Residual copies of deleted data may persist in routine backups for a limited period before being overwritten in the ordinary course.

11. Security

We use technical and organizational measures designed to protect personal information, including:

Encryption of data in transit using HTTPS.
Removal of embedded metadata from thumbnails on your device before any transmission.
One-way hashing of passwords (bcrypt); plaintext passwords are never stored. We additionally check new and changed passwords against publicly known data-breach databases using the Have I Been Pwned k-anonymity range API; your password is never transmitted, only a five-character prefix of its SHA-1 hash is sent, which alone cannot identify your password.
Storage of authentication and authorization tokens as cryptographic hashes rather than in plaintext, and an authorization design for connected AI assistants based on OAuth 2.1 with proof-key protection and token rotation.
Row-level security on our databases so that each account can access only its own data.
Rate limiting on sign-in and account-creation requests to defend against credential-stuffing and abuse.
Stateless processing of thumbnails, so that no image content is retained by our processing provider.

No method of transmission or storage is completely secure, and we cannot guarantee absolute security. If we become aware of a breach affecting your personal information, we will notify you and the relevant authorities within the timeframes and in the manner required by applicable law.

12. Your privacy rights

The rights available to you depend on where you live. To exercise any right, contact privacy@avlok.ai. We will not discriminate against you for exercising your rights. We may need to verify your identity before fulfilling a request, and where the law allows we may decline requests that are unfounded, excessive, or repetitive.

Note that some rights are subject to current product limitations. In particular, a self-service data export is not yet available; until it is, we will fulfill verified access and portability requests manually and within the periods required by law.

If you are in the EEA or the United Kingdom, you have the right to:

Access the personal information we hold about you and obtain a copy of it.
Rectify inaccurate or incomplete personal information.
Erase your personal information (“right to be forgotten”). You can delete your account and its associated data through Settings, and you may also contact us.
Restrict our processing in certain circumstances.
Object to processing based on our legitimate interests, and to object to direct marketing at any time.
Data portability, that is, to receive certain personal information in a structured, commonly used, machine-readable format.
Withdraw consent at any time where we rely on consent (for example, Cloud AI Access, marketing, or analytics), without affecting the lawfulness of processing before withdrawal.
Lodge a complaint with your local supervisory authority. We would, however, appreciate the chance to address your concerns first.

12.2 California (CCPA / CPRA)

If you are a California resident, you have the right to know, to access, to correct, and to delete the personal information we have collected about you; to opt out of the sale or sharing of your personal information; and to limit the use and disclosure of sensitive personal information. We will respond to verifiable requests within the period required by law (generally forty-five (45) days, extendable once where permitted).

No sale or sharing. We do not sell your personal information and we do not share it for cross-context behavioral advertising. Because we do not sell or share, the related opt-out has no application, although you may still exercise it as a matter of right.

Sensitive personal information. We do not use or disclose sensitive personal information for purposes that would trigger the right to limit beyond providing the Service you have requested and the permitted business purposes. Where you enable Cloud AI Access, the GPS coordinates that may be synced can constitute precise geolocation, which is treated as sensitive personal information; we use it only to provide location-based search within the Service.

The categories of personal information we collect, drawn from the categories enumerated under the CCPA/CPRA, are:

Statutory category	Examples in Avlok	Sold or shared?
Identifiers	Email address; first and last name; account identifier; country code; transient IP address; random analytics identifier	No
Customer records (Cal. Civ. Code 1798.80(e))	Name and account/billing details	No
Commercial information	Subscription and transaction records; marketing preference; optional referral source	No
Internet or network activity	Usage counts; anonymous interaction events	No
Geolocation data	For Cloud AI Access users, coarse location names and, where present, GPS coordinates of synced photos	No
Sensitive personal information	Account credentials; for Cloud AI Access users, precise geolocation (GPS coordinates)	No
Inferences and content-derived data	Content descriptions of your photos retained on our servers; for Cloud AI Access users, search representations and metadata stored on our servers	No

We collect these categories for the business purposes described in Section 5 and retain them as described in Section 10.

Authorized agents. You may use an authorized agent to submit a request on your behalf. We may require the agent to demonstrate authority and may ask you to verify your identity directly.

12.3 Other United States state privacy rights

If you are a resident of a United States state with a comprehensive privacy law (which currently includes, among others, Virginia, Colorado, Connecticut, Utah, Texas, Oregon, Montana, and a growing number of others), you may have rights to:

Confirm whether we process your personal information and access it.
Correct inaccuracies in your personal information.
Delete personal information we have collected about you.
Obtain a portable copy of personal information you have provided.
Opt out of targeted advertising, the sale of personal information, and certain profiling. We do not engage in targeted advertising, sell personal information, or conduct profiling that produces legal or similarly significant effects.

Where your state provides a right to appeal a decision on your request, you may appeal by contacting privacy@avlok.ai with the word “Appeal” in your message. If we deny your appeal, your state’s instructions for contacting its Attorney General will apply.

13. Marketing and other communications

We send two distinct categories of email.

Transactional email is sent in direct response to your actions or to inform you about your account. Examples include a welcome message when you register, a code when you request a password reset, and a confirmation when you delete your account. Because these are operational notifications rather than marketing, they cannot be unsubscribed from while your account is active; you can stop them only by deleting your account.

Marketing email consists of product announcements and similar non-transactional messages, and is opt-in only. At registration we present an unchecked option you may select to receive it. You can change your preference at any time in Settings, and you may unsubscribe using the link in any marketing message. We do not sell your email address or share it with third parties for their own marketing.

All email is delivered by our email provider (see Section 8), which receives your email address and, where present, your first name, and nothing else.

14. Cookies and similar technologies

The Avlok desktop application is not a website and does not use advertising cookies or third-party tracking cookies. It stores limited information locally on your device that is necessary to operate, such as your local search index and a sign-in session. The anonymous analytics identifier described in Section 15 is stored locally on your device.

15. Anonymous analytics in detail

If you do not opt out, Avlok sends anonymous usage events to our analytics provider. Each event carries:

An event name, such as “app launched” or “search executed.”
A small set of bucketed properties that record the shape of an action rather than its content. For example, we may record that a query contained a certain number of words, the number of results returned, the duration of an operation, and the operating-system platform. We do not record the text of your query.
A random identifier created on your first launch and stored locally. It is not linked to your account, your email, or any other identifier we hold, and deleting your local data resets it.

By technical design, the following are never sent to our analytics provider: your email address or name, the text of your search queries, file names or file paths, photo or video content, and your IP address. You may turn analytics off at any time in Settings; doing so stops new events from being sent and clears any events queued on your device. You may also ask us to delete the record associated with your random identifier. If we move to a self-hosted analytics solution in the future, we will update this policy and remove the third-party analytics provider as a sub-processor.

16. Automated decision-making and profiling

The Service uses automated processing to organize and search your library (for example, generating search representations and grouping photos). This processing supports the search functionality you request and does not produce legal effects concerning you or similarly significantly affect you within the meaning of Article 22 of the GDPR. We do not make automated decisions about creditworthiness, employment, or similar matters, and we do not engage in profiling for advertising.

17. Children’s privacy

Avlok is not directed to children. You must be at least 13 years old to create an account, or older where your jurisdiction requires. In the EEA and the United Kingdom, the minimum age for consent to information-society services may be higher (up to 16, depending on your member state); where a higher age applies in your jurisdiction, you must meet that age. We confirm age at registration and do not knowingly collect personal information from anyone under the applicable minimum age. If you believe a person under the applicable age has created an account, contact privacy@avlok.ai and we will delete it.

18. Changes to this policy

We may update this policy from time to time. When we do, we will update the version number and effective date above. For material changes, we will provide notice through the Service and, where appropriate, by email, and where required by law we will obtain your consent or ask you to accept the updated policy before you continue to use the Service.

19. Third-party attributions

This product includes GeoLite2 data created by MaxMind, available from https://www.maxmind.com. The GeoLite2-Country database is used within our own infrastructure to resolve a registration country code from an IP address; no IP data is sent to MaxMind or any other third party for this purpose.

20. Contact

For privacy questions, requests, or concerns, contact:

Avlok AI, LLC 502 West 7th St, STE 100, Erie, PA 16502-1333, United States privacy@avlok.ai

We aim to respond within five (5) business days.