Privacy Policy

Last updated: November 14, 2026

This Privacy Policy explains how Aqeeli LLC ("Aqeeli", "we", "us" or "our") collects, uses, shares and protects personal data when you use Context OS. That includes the website at https://contextos.dev, the application at https://app.contextos.dev, the browser extension, the MCP endpoint and our APIs (together, the "Service").

It also explains your rights under the EU and UK General Data Protection Regulation (GDPR), the Saudi Personal Data Protection Law (PDPL), and California privacy law.

1. Who we are and scope

The controller of personal data processed through the managed Context OS service is Aqeeli LLC, in formation as a Delaware limited liability company in the United States. Contact: privacy@contextos.dev.

Postal address: Aqeeli LLC (in formation) — Al-Sanabel District, Jeddah, Kingdom of Saudi Arabia

EU and UK residents can send privacy inquiries and data subject requests directly to our Data Protection & Privacy Office: privacy@citnexus.com or fadi@citnexus.com.

This policy does not apply to self-hosted deployments. When an organization runs Context OS on its own infrastructure under an enterprise license, that organization is the controller (or processor) of the data in its instance. Its own privacy notice applies.

When a team or business uses a shared Team vault, the organization that administers it decides what content is uploaded and who can access it. Members should also consult that organization's own policies.

2. Key points

  • The free browser audit runs entirely in your browser. Nothing you analyze there is uploaded to us.
  • Before upload, your browser redacts common secrets (such as API keys, passwords, IBANs and card numbers), email addresses and phone numbers from imported files, quick-start text and browser-extension captures. Our servers redact the same from voice-note transcripts and from what connected AI tools write. Notes you write in the editor or sync from a local folder, and your Vault Chat questions, are checked for secrets only: email addresses and phone numbers in them are kept. Nothing removes names, addresses, health data or other personal data (see Section 4).
  • On our managed cloud, sanitized text is processed in server memory by our AI provider to build your vault. Vault files are encrypted at rest. This is not zero-knowledge while a vault is being distilled or served to AI tools. An optional passphrase lock makes stored vault files zero-knowledge at rest.
  • We do not sell or share your personal data for advertising. Under the paid-tier terms we use, our AI providers do not use your API inputs or outputs to train their models.
  • You can export your vault at any time and delete your account in Settings → General.

3. Data we collect

3.1 Account data

  • Name, email address, password (stored as a hash), profile picture (if you upload one) and language preference.
  • Sign-in data from passkeys and two-factor authentication.
  • If you sign in with Google or GitHub, basic profile information from that provider, such as your name, email address and avatar.
  • Organization and team membership, roles and invitations.

3.2 Vault content

This is the content you choose to put into the Service:

  • AI chat exports, documents (PDF, DOCX, CSV, HTML and Notion exports) and pasted text;
  • voice notes and their transcripts;
  • conversations captured by the browser extension;
  • edits made by you or by connected AI agents; and
  • the Markdown files, routing maps, summaries, tags, quarantine proposals and change history generated from all of the above.

Vault content may contain personal data about you and about other people. That depends entirely on what you upload.

3.3 Payment data

Payments are handled by Stripe. We receive limited information from Stripe: your customer and subscription identifiers, the plan purchased, the amount, the payment status, and the country used for tax purposes. Your card details never reach our servers.

3.4 Technical and usage data

  • IP address, browser and device information, and request logs, used for security, abuse prevention and debugging.
  • Service usage records needed to run your plan: token usage, number of agent calls, tokens served, job status, last API token use, and rate-limit counters.
  • Identifiers or SHA-256 hashes of sources already processed, so the same conversation is not processed or charged twice. These records do not contain the content itself.
  • Product usage events, kept in our own database: steps such as running the public demo, finishing an audit, signing up, starting or finishing a build, asking a first Vault Chat question, connecting an AI tool, clicking an upgrade link and completing a checkout. Each event holds your account identifier (and your organization's, where relevant), or the random demo identifier described in Section 3.5 while you are signed out, plus a few coarse values such as a plan, a sign-in method or a rounded count. They never hold your email address, your name, file names or vault content. We use them in aggregate to improve the Service (legitimate interests, see Section 6). When you delete your account, they are no longer linked to you. We also send these events to PostHog (Section 8).

3.5 Cookies and similar technologies

We use only what the Service needs to work. These are all the cookies we set, and we do not show a cookie banner:

  • Session cookie (app): keeps you signed in, for up to 30 days.
  • Language cookie (NEXT_LOCALE, website and app): remembers your interface language.
  • Referral cookie (ctxos_ref): set only when you arrive through an affiliate or creator link, for up to 60 days, so that an account you create is credited to that partner. The first link you followed is kept.
  • Sidebar cookie (sidebar-collapsed, app): remembers whether you collapsed the sidebar.

We also use browser storage, which stays on your device:

  • Local storage: your light or dark theme; in the app, a random identifier (ctxos_aid) that counts the steps of the public demo in our own statistics while you are signed out.
  • IndexedDB in the app: remembers a linked local folder, and keeps the local .env.example values described in Section 4.
  • Extension storage: holds your extension settings, your token and a cached copy of your vault's AGENTS.md.

At the date of this policy, no third-party analytics or advertising scripts are loaded on the Service. If we add analytics in the future, they will run only after you consent where the law requires consent, and we will update this policy first.

3.6 Contact form and newsletter

If you write to us through the contact form on our website, we store your name, email address, company (if you give one), topic, message and interface language, and we forward the request to our team's inbox through our email provider. We use it only to answer your request and follow up on it. Legal basis: legitimate interests in answering the people who contact us, and steps taken at your request before a contract.

If you subscribe to our newsletter, we store your email address and interface language and add the address to our mailing list at Resend (see Section 8). Legal basis: consent, which you can withdraw at any time.

These records are not linked to an account, so deleting your account does not remove them. To unsubscribe, use the unsubscribe link in any newsletter email: it deletes your subscription and removes your address from the mailing list at once. You can also email privacy@contextos.dev to unsubscribe or to have a contact request or subscription deleted. We then delete the record and remove your address from the mailing list.

4. What happens in your browser

Before an imported file, quick-start text or a browser-extension capture is uploaded, the Service sanitizes it in your browser. This step removes noise and runs our local scrubber. The scrubber looks for patterns and replaces what it finds with placeholders:

  • Secrets: private keys; passwords in database connection strings; API keys and tokens from common providers (such as OpenAI, Anthropic, Google, Stripe, GitHub, AWS, Supabase, Slack and SendGrid); JSON Web Tokens; bearer and basic-auth credentials; secret-looking environment variable assignments; values that follow words such as "password", "passphrase", "secret" or "API key"; payment card numbers (checksum-validated); IBANs (checksum-validated); and other high-entropy strings that look like keys. These become placeholders such as [REDACTED_SECRET_1].
  • Email addresses become [REDACTED_EMAIL_n], and phone numbers become [REDACTED_PHONE_n].

The real values of key-like secrets are kept only on your device. They are stored in your browser and written to a local .env.example file in your linked folder or in your ZIP download. They are never uploaded and are excluded from bring-your-own-storage sync. Our servers run the same scrubber again on uploads and extension captures as an extra safeguard, and they run it on voice-note transcripts and on what connected AI tools write before anything is stored.

Text you write yourself. Notes you write in the editor or sync from a local folder, and your Vault Chat questions, are checked for the secrets above only, because they often contain contact details on purpose. Email addresses and phone numbers in them are kept.

  • Local folder sync replaces secrets in your browser before upload, and the files on your computer keep the real values. Files sent to our sync API with a secret in them are not stored.
  • The editor checks every save on our servers. A save that contains a secret is refused until you confirm, and the secret is then stored as a placeholder.
  • Vault Chat removes secrets from your questions on our servers before they reach our AI provider.

Important limits. The scrubber is pattern-based. It can miss secrets, and it sometimes redacts harmless text. It does not remove names, postal addresses, identification numbers, health information, financial details other than card numbers and IBANs, or other personal data about you or third parties. You are responsible for what you upload. Please remove anything you do not want processed.

5. How we process vault content

Distillation. When you build or update a vault on our managed cloud, the sanitized text is uploaded, staged briefly in our database and processed in server memory. It is sent to our AI provider (see Section 8) to extract facts and write your vault files. Staged upload text is deleted when the distillation job ends. If your token allowance runs out during a build, the sources that were not distilled stay staged, encrypted with your vault's key, for up to 7 days, so a purchase or "Distill the rest" can finish the build without a second upload; they are deleted after 7 days, when you lock the vault with a passphrase, or when you delete the vault. Our cleanup job also deletes staged text from abandoned, failed or completed jobs 24 hours after the job last changed. Our background-job provider keeps intermediate step results, which can include distilled content, as part of its run history.

Encryption at rest. Vault files and quarantine diffs are encrypted at rest with a per-vault AES-256-GCM data key. That key is itself encrypted ("wrapped") with a master key held only on our servers.

Not zero-knowledge in active use. Distillation, the MCP endpoint, the browser extension, Vault Chat, voice notes and folder reorganization all need readable, secret-scrubbed text. That text is decrypted and processed in server memory while those features run. So the managed cloud is not zero-knowledge while a vault is being distilled or served.

Optional passphrase lock. If you lock a vault with a passphrase, your browser derives a key (PBKDF2-SHA-256) and encrypts every vault file before storage. We store only the parameters needed to verify the key, never the passphrase or the key. While a vault is locked, its stored files are zero-knowledge at rest. AI, MCP, extension, distillation and voice features are paused until you unlock it in your browser.

Vault Chat. Your questions and the relevant vault files are sent to the AI provider to generate an answer. We do not keep a server-side history of Vault Chat conversations.

Voice notes. If your browser supports speech recognition, it transcribes your voice note. Some browsers, such as Google Chrome, process that audio on the browser vendor's servers under the vendor's terms. If you use our server transcription fallback, the audio is sent to OpenAI for transcription. The transcript is then scrubbed and routed into your vault.

Browser extension. On supported AI chat sites (ChatGPT, Claude, Gemini and Grok), the extension can insert your vault context into the chat. It can also capture the conversation after it pauses, scrub it locally and send it to your vault. You can turn auto-injection and capture off in the extension settings. The extension does not read other websites unless you grant it permission for a self-hosted server address.

MCP and share links. AI tools you connect with a vault API token, and guests you invite with an expiring share link, can read the parts of your vault that the token's role or link allows. We store only salted hashes of tokens and share links.

Bring your own storage. If you connect GitHub, another git host or an S3-compatible bucket, we push your vault files there using credentials that we store encrypted. That provider then processes the files under its own terms. Distillation still runs on our servers.

Bring your own key. If you connect your own AI provider key, AI calls for your vault go to your provider account under that provider's terms. We store your key encrypted.

6. How we use personal data and our legal bases

  • Provide the Service (create your account, build and serve your vaults, sync, extension, MCP, Vault Chat, support). Legal basis: performance of our contract with you.
  • Billing, token quotas and fraud prevention. Legal basis: contract; legitimate interests in preventing abuse; and legal obligations (tax and accounting).
  • Security, debugging and abuse prevention (logs, rate limits, enforcement of our Acceptable Use Policy). Legal basis: legitimate interests in keeping the Service secure.
  • Transactional email (verification, password reset, invitations, billing and service notices). Legal basis: contract and legitimate interests.
  • Contact requests and the newsletter (Section 3.6). Legal basis: legitimate interests and steps taken at your request before a contract for contact requests; consent for the newsletter.
  • Improving the Service using aggregated usage statistics, such as token totals and feature use. Legal basis: legitimate interests. We do not use your vault content for this.
  • Analytics cookies, if we introduce them. Legal basis: consent, which you can withdraw at any time.
  • Complying with law and responding to lawful requests. Legal basis: legal obligation.

We do not use your vault content to train AI models, and we do not sell your personal data.

7. Where your data is shared

We share personal data only:

  • with the subprocessors listed in Section 8, to run the Service;
  • with people you choose to share with, such as team members, share-link guests, connected AI tools and your own storage providers;
  • if the law requires it, or to protect the rights, safety or property of our users, the public or us; and
  • with a successor in a merger, acquisition or sale of assets. That successor will remain bound by this policy, or we will notify you.

8. Subprocessors

Each subprocessor below is listed with its purpose, the data it processes and its location:

  • Google (Gemini API)
    • Purpose: AI processing for distillation, routing, Vault Chat and voice-note routing.
    • Data: sanitized vault text and prompts.
    • Location: United States and other Google locations.
    • Note: we use the paid tier, under which Google does not use API inputs or outputs to train its models.
  • OpenAI
    • Purpose: server-side voice transcription, only when you use it.
    • Data: voice-note audio.
    • Location: United States.
    • Note: we use the paid API. OpenAI does not train on API data by default.
  • Neon
    • Purpose: managed PostgreSQL database.
    • Data: account data, encrypted vault files, staged uploads and usage records.
    • Location: United States.
  • Vercel
    • Purpose: application hosting and serverless compute.
    • Data: all data processed by the app in transit and in memory, and request logs.
    • Location: United States, with a global edge network.
  • Inngest
    • Purpose: background job orchestration (distillation, cleanup, storage sync).
    • Data: job identifiers and status, plus intermediate step results that can include distilled content.
    • Location: United States.
  • Resend
    • Purpose: transactional email, and our newsletter mailing list.
    • Data: email address, name and email content; for newsletter subscribers, the email address.
    • Location: United States.
  • PostHog
    • Purpose: product analytics of our conversion funnel (steps such as sign-up, build and checkout).
    • Data: your account identifier or the random demo identifier, the event name and a few coarse values such as a plan or a rounded count; never your email address, name, file names or vault content. No person profiles are created.
    • Location: United States (or the European Union if we use PostHog's EU region).
  • Stripe
    • Purpose: payments, subscriptions and tax.
    • Data: billing details, payment method and transaction history. Card data goes directly to Stripe.
    • Location: United States and other Stripe locations.

If you use bring-your-own-key or bring-your-own-storage, the providers you choose are not our subprocessors. They act under your own agreement with them. We will update this list before adding a new subprocessor that processes vault content.

Team and Enterprise organizations: our processor terms for team vaults, including this list, security measures and how to request a signed copy, are in the Data Processing Addendum.

9. International transfers

We are based in the United States, and our subprocessors mainly process data there. If you are in the European Economic Area, the United Kingdom, Switzerland, Saudi Arabia or another country with transfer rules, your data is transferred to the United States. We rely on:

  • the EU-U.S. Data Privacy Framework (and its UK and Swiss extensions) where the recipient is certified; or
  • the European Commission's Standard Contractual Clauses (with the UK addendum where relevant) and supplementary measures, such as encryption in transit and at rest.

For Saudi Arabia, see Section 13. You can request more information about these safeguards at privacy@contextos.dev.

10. How long we keep data

  • Account data: while your account exists. It is deleted when you delete your account.
  • Vault content: until you delete it, delete the vault, or delete your account. Drafts in a vault's 99_ temporary folder are purged automatically after 7 days.
  • Staged upload text: deleted when the distillation job ends. Leftovers are purged 24 hours after the job last changed. Sources a budget-limited build could not distill are kept encrypted for up to 7 days, then deleted.
  • Public profile: if you publish a vault's SOUL.md as a public profile, anyone with the link can see it until you unpublish it.
  • Contact requests and newsletter subscriptions: until you unsubscribe or ask us to delete them (Section 3.6).
  • Processed-source identifiers and hashes: while the vault exists.
  • Sessions: expire after 30 days at most.
  • Payment records: our database copy is deleted with your account. Stripe keeps transaction and invoice records as required by tax and financial laws.
  • Logs and backups: kept for limited periods under our providers' standard retention. Deleted data may remain in encrypted backups until those backups expire.

11. Security

We use technical and organizational measures to protect personal data:

  • encryption in transit (TLS) and per-vault AES-256-GCM encryption of vault files at rest;
  • encrypted storage of AI keys and storage credentials, and hashed API tokens and share links;
  • role-based access to folders, rate limiting, and spend controls; and
  • restricted administrative access.

No system is completely secure. If a personal data breach affects you, we will notify you and the competent authorities as the law requires.

12. Your rights (EU, UK and similar laws)

Depending on where you live, you have the right to:

  • access your personal data and receive a copy;
  • rectify inaccurate data;
  • erase your data;
  • restrict processing;
  • data portability, that is, to receive your data in a structured, machine-readable format (your vault exports as Markdown or ZIP);
  • object to processing based on legitimate interests;
  • withdraw consent at any time, where processing is based on consent; and
  • lodge a complaint with a data protection supervisory authority, in particular in your country of residence or work. In the UK, that is the Information Commissioner's Office.

To exercise your rights, email privacy@contextos.dev from your account email address. We may need to verify your identity. We respond within 30 days. Where the law allows us to extend that period, we will tell you why.

13. Saudi Arabia (PDPL)

If you are in the Kingdom of Saudi Arabia, the Personal Data Protection Law and its Implementing Regulations give you the following rights:

  • the right to be informed about how and why your personal data is collected and processed (this policy);
  • the right to access your personal data and obtain a copy in a readable, clear format;
  • the right to request correction, completion or updating of your personal data;
  • the right to request destruction of your personal data when it is no longer needed, subject to legal retention requirements; and
  • the right to withdraw consent to processing that is based on consent.

Cross-border transfer. Your personal data is transferred to and processed in the United States by us and by the subprocessors in Section 8. We transfer it in order to provide the Service you request, and only the minimum data necessary. We use contractual and technical safeguards, such as encryption, consistent with the PDPL's rules on transfers outside the Kingdom.

For any PDPL request, or any matter you would like to raise relating to the Saudi Data and Artificial Intelligence Authority (SDAIA), contact privacy@contextos.dev. You may also file a complaint with the competent authority under the PDPL.

14. California (CCPA/CPRA)

California residents have the right to know what personal information we collect, use and disclose. They also have the right to request deletion and correction, and the right not to be discriminated against for exercising these rights. The categories we collect are described in Section 3, and we collect them for the purposes in Section 6.

We do not sell or share personal information, as those terms are defined under California law. We do not use or disclose sensitive personal information for purposes that would give you a right to limit it. To make a request, email privacy@contextos.dev. You may use an authorized agent.

15. Children

The Service is not directed to children under 16, or under the age of digital consent in your country if that is higher. We do not knowingly collect their personal data. If you believe a child has given us personal data, contact privacy@contextos.dev and we will delete it.

16. Exporting and deleting your data

  • Export: you can export your vault at any time as Markdown files or a ZIP archive, or sync it to a local folder or your own storage.
  • Delete content: you can delete individual files, folders or whole vaults in the app.
  • Delete your account: go to Settings → General → Delete account. Deleting your account cancels any active subscription. It also permanently deletes, from our database, your personal vaults with their files and folders, your API tokens and share links, your ingestion jobs and your purchase records. Vaults that belong to an organization are deleted when that organization is deleted.
  • By email: you can also send any request to privacy@contextos.dev. We respond within 30 days.

17. Changes to this policy

We may update this policy from time to time. If we make material changes, we will notify you by email or in the app before they take effect. The "Last updated" date shows the latest version.

18. Contact

Aqeeli LLC, in formation as a Delaware limited liability company, United States.

EU and UK residents can send privacy inquiries and data subject requests directly to our Data Protection & Privacy Office: privacy@citnexus.com or fadi@citnexus.com.