Legal
Privacy Policy
How Relay Port handles Instagram data: what we collect, why, how long we keep it, and how to delete it.
Last updated: November 20, 2026
About this policy
Relay Port, operated by Bypass Inc. (バイパス株式会社)
Effective date: 2026-11-21
This is the privacy policy for Relay Port, our social media publishing and scheduling API, dashboard and website at relayport.dev. It explains what personal data we collect, why, who we share it with, how long we keep it, and how you can have it deleted.
Relay Port is a business product. Our customers are developers, AI-agent builders and agencies. They connect social media accounts to Relay Port so that posts can be published and scheduled to those accounts. In this release, the only platform is Instagram.
1. Who we are
- Company: Bypass Inc. (バイパス株式会社), a company incorporated in Japan
- Address: Park Heights Ikebukuro 1105, 1-17-11 Higashi-Ikebukuro, Toshima-ku, Tokyo 170-0013, Japan
- Representative: Yousef Kenji, Representative Director (代表取締役 剣地 ユセフ)
- Privacy contact: privacy@relayport.dev
- Data protection officer or EU/UK representative: contact privacy@relayport.dev, Bypass Inc., at the address above
In this policy, "we", "us" and "Relay Port" mean Bypass Inc., the operator. "Customer" means the person or company that holds a Relay Port workspace. "You" means whoever the policy is talking to in context: a Customer's team member, or a person who connects an Instagram account to a Customer's workspace.
Our role
- Controller. For the data we need to run our own service (sign-up details, billing, security logs, support), we decide why and how it is used. We are the controller (GDPR) or the business operator handling personal information (APPI, Japan).
- Processor. For the content a Customer publishes through Relay Port, and for the data of the Instagram accounts a Customer connects, we act on the Customer's instructions. The Customer decides what to publish and to which accounts. We are a processor (GDPR) or an entrusted party (APPI). Our Data Processing Addendum covers this.
2. What we collect, and where it comes from
2.1 When you sign up and use the dashboard
- Email address, name, password (stored only as a salted and peppered hash). Why we need it: Create and secure your account
- If you sign in with Google: your Google account identifier, email address and name. Why we need it: Sign you in
- Workspace details: workspace name, brands, members, roles, invitations. Why we need it: Run your workspace
- API keys (we store only a one-way hash, and show you the key once) and key scopes. Why we need it: Authenticate your API access
- Webhook URLs a workspace registers through the API, each with an optional description, the event types and the brands or accounts it covers, whether it is for test or live data, and its signing secret (encrypted, and shown only when it is created or replaced). Why we need it: Keep the addresses and signing secrets you set up for receiving events
- Sessions and sign-in records. Why we need it: Keep you signed in and detect misuse
- AI clients you authorize through OAuth (for example Claude or ChatGPT): the client's identifier and name, its registration details (name and redirect addresses), the workspace, brands and permissions you granted, who granted them, when, and when they were revoked. Access tokens, refresh tokens and authorization codes are stored only as one-way hashes. When a client identifies itself with a client metadata document, we fetch that document from the client's own website. Why we need it: Let the AI clients you approve act in your workspace within the permissions you chose, and let your team see and revoke them
- Your messages to us (support requests, feedback, emails to our addresses). Why we need it: Reply and improve the service
- "I'm interested" clicks on a platform or module that is not available yet: which item you clicked, your user ID, your workspace ID and the time. Why we need it: Count demand so we know what to build next. Our staff tools show only totals per item and per release, never who clicked
2.2 When an Instagram account is connected
A workspace member can connect an Instagram professional (Business or Creator) account. This uses Instagram API with Unified Login (Facebook Login for Business), a Meta product. You log in with Instagram or Facebook, choose the business portfolio and the Instagram accounts to share, and Meta shows you its own consent screen where you choose whether to allow access.
We ask for two permissions:
- instagram_basic, to read the basic profile of the connected account.
- instagram_content_publish, to publish content to the connected account.
From Meta we receive and store:
- Instagram account ID, username and profile picture URL. Why: Show you which account is connected, and address it correctly when publishing
- The access token that Meta issues for your business portfolio (and its expiry, if it has one). Why: Publish on your behalf. Tokens are encrypted at rest (AES-256-GCM). See section 9
- Permissions you granted. Why: Know what the connection can do and tell you when you need to reconnect
- The media and captions you publish, the resulting post IDs and permalinks, and the publishing status (including errors Meta returns). Why: Publish, schedule, show you what happened and let you open the post on Instagram
- Account limits that Meta reports (for example the remaining daily publishing quota). Why: Stay inside Instagram's limits
What we do not collect in this release: direct messages, comments, insights (statistics), audience data, or ads data. If we add any of these later, we will do it only after updating this policy, and (where Meta requires it) after Meta's review and your fresh consent.
Connect links. A Customer can send the holder of an Instagram account a connect link, so that the account holder connects their own account to the Customer's workspace without signing in to Relay Port. A connect link records:
- Who created it: the workspace member or the workspace's API key.
- The workspace and the brand that the account will be connected to, and the platforms the link offers.
- Whether it connects a new account or reconnects one named account that is already connected.
- The address and reference the Customer asked us to send the account holder back to, if any.
- When it expires (at most 30 days after it was created), and when it was used or revoked.
We keep only a one-way hash of the link's secret code, never the code itself, and the page that opens the link sends no referrer and is kept out of search engines. The person who opens the link sees the brand's name, logo and colour and the platforms offered, nothing else about the workspace. When they connect, we store the connected account's public profile (account ID, username and profile picture URL) and its access token as described above, in the Customer's workspace and in that brand. From then on, the account is visible and usable like any other account the Customer connected (see section 5): to the workspace members with the right role and to API keys that the workspace has authorised with the right scope. The account holder does not get access to the workspace. A link stops working when it expires, after one successful connection, when the Customer revokes it, or when the member or API key that created it loses access to the workspace or the brand.
2.3 Content you upload or schedule
When you upload images or videos, we store the files in private cloud storage. You can send us a file directly, or ask for an upload link: an address in our storage, signed for the file's declared type and size, that expires after 15 minutes. We check the file when you tell us the upload is complete.
You can also give us the web address of an image or video to import. We then fetch that one address from our servers, only over HTTPS. We refuse addresses that name a private, loopback or link-local network, localhost or our own domain, we never follow a redirect, we accept only JPEG images up to 8 MB and MP4 videos up to 50 MB, and we give up after 60 seconds. We keep the address with the file's record. A file that fails our checks is deleted from storage at once, and its record keeps the reason.
You can add alt text (a description of an image for people who cannot see it) to an image. We store it with the image and send it to Instagram with the image when it is published as a single-image post.
When Instagram needs to fetch a file in order to publish it, we serve it from a signed, expiring link. We also store your captions, schedule and per-target publishing records ("receipts"), including the evidence we gathered when Instagram's answer was unclear.
Content you upload is yours. We do not use it for our own purposes other than providing the service to you.
2.4 Payments
AI credits. The workspace owner can buy AI credit packs through Stripe Checkout. If the workspace has no Stripe customer yet, we create one first, as described below for paid plans. For each purchase we send Stripe the workspace ID, the credit type and our purchase ID, and we store the pack, the amount, the purchase status, the Stripe checkout session and payment IDs and when it was paid. We also keep a ledger of the workspace's credits: each purchase, use, return, expiry and refund, with its amount, the purchase or request it relates to, and the expiry date. When Stripe tells us that a payment was refunded, disputed or flagged as possibly fraudulent, or that a refund's status changed, we fetch from Stripe that payment's amount, its refunds and their status, and the purchase details we attached to the payment (workspace ID, credit type and purchase ID); a dispute or a fraud warning is recorded on the workspace.
Paid plans are billed by Stripe. When you subscribe, Stripe collects your payment details, billing address and tax ID on its hosted checkout page. We do not receive or store full card numbers. We store your Stripe customer ID, plan, subscription status, invoice records and the number of connected accounts billed. Billing is not enabled until paid plans launch.
When the workspace owner first starts a paid plan, we create a Stripe customer for the workspace and send Stripe the owner's email address and name and the workspace ID. Stripe processes this data, and the details you give on its checkout page and in its customer portal, as our payment processor. When the owner opens the Billing page, we fetch the workspace's most recent invoices from Stripe to show them there. When a payment fails or the bank asks for a payment to be confirmed, we email the workspace owners a link to the invoice on Stripe. When a workspace reaches 80% or 100% of a plan limit, or its AI credits run low, we email the workspace owners and add a notice to their in-app notifications; an owner can turn these emails off by category in notification settings. Among workspace members, only the owner can see billing details and invoices in Relay Port. Our staff with system administration rights can also see a workspace's billing details and invoices, and the records we keep of the payment events Stripe sends us, to provide support and resolve billing problems. Those records hold Stripe identifiers, the event type and its processing status, not names, email addresses or postal addresses. When staff need an event's full details, they fetch them from Stripe at that moment; we do not store them.
2.5 Logs and technical data
- Request logs for the API and dashboard: IP address, user agent, time, endpoint and result.
- A workspace request log that you can read. For each request made with an API key, and each request a workspace member sends to change something in the workspace, it records the time, method, path (without the query string), response status, response time, request ID, live or test mode, and the API key's prefix or the member's user ID. It holds no IP address, headers or request bodies. Workspace members with the developer role or above can read it; API keys cannot.
- A delivery log of each publishing attempt, with error codes (including Meta's fbtrace_id, which Meta asks us to quote in support requests).
- An event log of what happened to your posts, connected accounts and uploaded media (for example "post published" or "account needs reconnecting"). Each entry keeps a copy of the post or account details at that moment, such as the caption and the account username. Workspace members with the developer role or above, and API keys allowed to read events, can read it through the API.
- A webhook delivery log: for each event we send to a webhook URL, one entry per attempt with its time, whether it succeeded, the HTTP status the receiving server answered, the first 2,048 bytes of that server's response and how long it took. Workspace members with the developer role or above, and API keys allowed to read webhooks, can read it through the API.
- Rate-limit and abuse-detection counters.
We use these to run, secure and debug the service, and to show you what happened to each post.
2.6 Website and cookies
relayport.dev and the dashboard use cookies that are strictly necessary (for example, to keep you signed in). We use no analytics or advertising cookies. If we add analytics, we will ask for consent first. Where the law requires consent, we ask first and give you a "reject all" choice.
3. Why we use your data, and our legal bases (GDPR)
- Provide the service: accounts, publishing, scheduling, receipts, API. Data: Sections 2.1 to 2.3. Legal basis: Performance of a contract (Art. 6(1)(b)). For a connected account holder who is not our Customer, our Customer's and our legitimate interests (Art. 6(1)(f)) in providing the publishing service the account holder asked for
- Billing, tax, accounting. Data: 2.4. Legal basis: Contract; legal obligation (Art. 6(1)(c))
- Security, abuse prevention, fraud, debugging. Data: 2.1, 2.5. Legal basis: Legitimate interests (Art. 6(1)(f)) in keeping the service secure
- Customer support. Data: 2.1, 2.5. Legal basis: Contract; legitimate interests
- Deciding what to build next from "I'm interested" clicks. Data: 2.1. Legal basis: Legitimate interests (Art. 6(1)(f)) in building what our customers ask for
- Product notices (changes, incidents, billing, reconnect alerts). Data: 2.1. Legal basis: Contract; legitimate interests
- Optional analytics and non-essential cookies. Data: 2.6. Legal basis: Consent (Art. 6(1)(a))
- Complying with law and Meta's platform requirements. Data: All. Legal basis: Legal obligation; legitimate interests
We do not sell personal data. We do not use data we receive from Meta for advertising, for profiling you, or to build a profile of you or your audience. We do not use data from Meta, or the content you publish, to train AI models.
4. Purposes of use under Japan's APPI
For personal information handled under Japan's Act on the Protection of Personal Information, the purposes of use are: (1) providing and operating Relay Port; (2) identity verification and authentication; (3) billing and payment processing; (4) customer support and communication; (5) security, abuse prevention and fraud detection; (6) improving the service in a way that does not identify individuals; (7) complying with laws. We will not use personal information beyond these purposes without your consent unless the law allows it.
5. Who we share data with
Inside your workspace. Data about a connected Instagram account is visible only to the Customer workspace that connected it, to the workspace members with the right role, and to API keys that the workspace has authorised with the right scope. We never show one Customer's connected-account data to another Customer.
Instagram (Meta). When you publish, we send your media and caption to Instagram, because you asked us to. Meta handles that data under its own terms and privacy policy.
Your webhook URLs. When a workspace registers a webhook URL, we send that URL the events it chose, as signed HTTPS requests. An event carries the same details as the event log (for example a post with its caption and publishing status, or a connected account's username). If a delivery fails we try again, up to 11 attempts over about three days, and a workspace member can send an event again or send a test event. The server at that URL is run by the Customer or by someone the Customer chose, and what happens to the data there is up to them, not us.
Sub-processors. We use a small number of companies to run Relay Port. Each handles data only on our instructions. The current list, what each does and where, is at relayport.dev/legal/subprocessors:
- Cloudflare: hosting, database, file storage, queues, email delivery and email routing.
- Stripe: payments and billing, once paid plans are enabled.
- Google: "Sign in with Google", if you choose it.
Staff. A small number of our staff can access production data. Access is limited by role. Staff with system administration rights can see billing details, invoices and payment event details as described in section 2.4.
Your own agents. If you connect your own AI agent to Relay Port, what your agent does with data it reads from Relay Port is governed by your agreement with that agent's provider, not by us. An AI client you authorize through OAuth receives data from one workspace only, limited to the brands and permissions you granted and to the role you hold at the time of each request. It stops receiving data as soon as the authorization is revoked, you leave the workspace or the workspace is suspended.
Legal and corporate. We may disclose data if the law requires it, to protect our rights or the safety of others, or in a merger or sale of the business. In a sale, we would require the buyer to honour this policy.
We do not sell or rent personal data, and we do not give it to data brokers.
6. International transfers
Bypass Inc. is in Japan. Relay Port runs on Cloudflare's global network, so data may be stored and processed in Japan, the United States and other countries where Cloudflare and our other sub-processors operate. We store data on Cloudflare's network (databases and file storage); it may be stored outside your country.
- From the EU, EEA, UK and Switzerland: we rely on the safeguards each provider offers, such as the EU Standard Contractual Clauses, the UK addendum, and where available the EU-U.S. Data Privacy Framework. Japan has an EU adequacy decision for personal data, which covers transfers to us as a Japanese business operator within its scope.
- From Japan (APPI Art. 28): when personal information is provided to a company in a foreign country, we ensure it has a system that meets the standards the law requires, or we obtain consent. Information about the foreign countries concerned and each provider's protection measures is available on request at privacy@relayport.dev.
7. How long we keep data
- Instagram access tokens. How long: Deleted immediately when the account is disconnected or when we process a data-deletion request
- Instagram profile data (account ID, username, profile picture URL). How long: While the account stays connected. After disconnect, kept 30 days so you can reconnect without losing history, then deleted automatically. If you use Delete my data or ask us by email, it is deleted without that wait
- Post records and receipts (captions, schedule, status, post IDs, permalinks). How long: While the account is connected and the workspace is active. Removed with the account's data after the period above. A post's status timeline is part of its delivery log and is kept as described next
- Uploaded media files. How long: Until you delete them or use Delete my data. Files that a pending post uses cannot be deleted until it is done. Signed delivery links to Instagram's fetchers expire after 24 hours, and upload links after 15 minutes. The web address a file was imported from and an image's alt text are kept with the file and erased when you delete it. A file that fails our checks is removed from storage at once; its record keeps the reason until you delete it. An upload you never complete is removed after 24 hours
- Delivery logs. How long: The log retention of the workspace's plan: 7 days on Free, 30 days on Pro, 90 days on Business and Enterprise, unless your agreement with us sets another period. Deleted sooner with the post, and with Delete my data
- Workspace request log. How long: The log retention of the workspace's plan, as for delivery logs, then deleted automatically
- Event log. How long: While the workspace is active. With the account's data (after the period above, or at once for a deletion request), entries about that account and its posts are deleted, and entries about a post it shared with other accounts keep only the other accounts' details. Every entry is deleted with Delete my data
- Webhook delivery log (the receiving server's status and the start of its response). How long: While the workspace is active. Entries for an account's events are deleted with that account's data; entries for a post the account shared with other accounts stay, including the start of the receiving server's response. Every entry is deleted with Delete my data or when the workspace is deleted. Deleting a webhook URL keeps its entries with the workspace's records
- Connect links (creator, workspace, brand, platforms, expiry, and when they were used or revoked). How long: While the workspace exists. A link that names an account is deleted with that account's data, and every link is deleted with Delete my data
- Webhook URLs and signing secrets. How long: Until you delete the URL or the workspace. Deleting a URL erases its signing secrets at once and keeps its address, description and event choices with the workspace's records until the workspace is deleted. A signing secret you replace stays valid for the overlap you choose (24 hours by default, at most 7 days) and is erased at the next replacement. When a connected account's data is deleted, the account is taken off each URL that also covers other accounts, and a URL limited to that account alone is turned off. Delete my data turns off every URL limited to particular accounts
- Duplicate-guard records. How long: Short-lived, up to 30 days
- Security and request logs (IP, user agent). How long: Up to 90 days
- AI client authorizations. How long: An access token works for 1 hour. Refresh tokens stop working 30 days after the client was authorized, or at once when the authorization is revoked; presenting a refresh token that was replaced more than a minute earlier revokes the authorization. A one-way hash of each refresh token we issue is kept for 30 days to detect that reuse. The record of the authorization (client, permissions, brands, who granted it, when, and when it was revoked) is kept while the workspace exists
- AI client registrations made through dynamic registration. How long: 90 days, renewed while the client keeps using them
- Account and workspace details. How long: Until you delete your account; then deleted within 30 days
- "I'm interested" clicks. How long: Until you delete your account, which deletes yours. Delete my data, or deleting the workspace, deletes every click made in that workspace
- Billing and tax records. How long: Up to 7 years after the transaction, as Japanese tax and accounting rules require
- Credit purchases and the credit ledger. How long: While the workspace exists; deleted with the workspace. Stripe keeps its own records of the payments and invoices
- Support emails. How long: Up to 24 months after the last message
- Backups. How long: Deleted data leaves our backups within 35 days of deletion
- Record of a data-deletion request (code, date, outcome; no content). How long: Kept so we can answer status look-ups
When we delete data, we remove it from the databases and file storage that serve the product. Where a law requires us to keep something, we keep only what is required.
8. Deleting your data
You can ask us to delete your data at any time:
- In the dashboard: open Settings and use Delete my data, which deletes the content of your whole workspace (every connected account, posts, delivery logs, uploaded media and "I'm interested" clicks) and keeps your sign-in, or use Disconnect on one account, whose data is deleted automatically 30 days later.
- By email: privacy@relayport.dev.
- In Meta: if you remove Relay Port in Meta, Meta sends us a removal request through its data-deletion callback. We receive and record it. To have your data deleted, use one of the two options above.
Full instructions, what is deleted, timelines and the status look-up are at https://www.relayport.dev/data-deletion. Disconnecting an account deletes its tokens immediately.
9. Security
- Instagram tokens are encrypted at rest with AES-256-GCM, with versioned keys, stored apart from other data, and never returned by our API. Disconnecting deletes them.
- API keys are stored only as hashes and are shown once.
- Signing secrets for webhook URLs are encrypted at rest with AES-256-GCM and shown only once, when they are created or replaced.
- OAuth access tokens, refresh tokens and authorization codes for AI clients are stored only as hashes, and every AI client, whether or not it also has a client secret or signing key, must prove it started the sign-in (PKCE with S256) before it receives a token.
- Events go to webhook URLs only over HTTPS, signed with the URL's signing secret. Before every request we check again that the URL is a public HTTPS address, we never follow a redirect, and we stop waiting for an answer after 10 seconds.
- Passwords are stored as salted, peppered hashes. Sessions are short-lived and revocable.
- Traffic uses HTTPS. Each workspace's data is isolated in its own scope in the database, and every query is limited to that workspace.
- Media is served only through signed, expiring links, and upload links are signed for one file's type and size.
- Files are imported from web addresses only over HTTPS. We refuse addresses that name a private, loopback or link-local network, localhost or our own domain, we never follow a redirect, and we stop reading at the size limit or after 60 seconds.
- Webhooks from Meta are verified by signature.
No system is perfectly secure. If we learn of a breach affecting your personal data, we will tell you and the authorities as the law requires. Report a vulnerability to security@relayport.dev.
10. Your rights
Everyone. You can ask us to access, correct or delete your data, or to tell us your concerns, at privacy@relayport.dev.
EU, EEA, UK and Switzerland (GDPR). You have the rights of access, rectification, erasure, restriction, portability and objection, the right to withdraw consent at any time, and the right to complain to your supervisory authority. We answer within one month.
Japan (APPI). You can request notice of purposes of use, disclosure, correction, addition, deletion, cessation of use, and cessation of third-party provision of your retained personal data, and disclosure of records of third-party provision. Contact: privacy@relayport.dev. There is no fee for these requests. You can also contact the Personal Information Protection Commission (個人情報保護委員会).
If you are a Customer's end user or an account holder. If a Customer connected your Instagram account, we process that data for them. You can also contact us directly, and we will help, or pass your request to the Customer where needed.
We may need to verify who you are before acting on a request.
11. Children
Relay Port is a business service for adults. It is not directed to children, and we do not knowingly collect personal data from anyone under 18. Instagram has its own minimum age. If you think a child has given us personal data, email privacy@relayport.dev and we will delete it.
12. Changes to this policy
We will post any change on this page and update the date at the top. For material changes, we will email the workspace owners at least 30 days before they take effect. If we add a new Meta permission or a new kind of data from Instagram, we will update this policy first and ask you to reconnect and consent.
13. Contact
- Privacy and data requests: privacy@relayport.dev
- Security reports: security@relayport.dev
- Support: support@relayport.dev
- Billing: billing@relayport.dev
- Post: Bypass Inc., Park Heights Ikebukuro 1105, 1-17-11 Higashi-Ikebukuro, Toshima-ku, Tokyo 170-0013, Japan, attention Yousef Kenji, Representative Director (代表取締役 剣地 ユセフ)
Instagram and Meta are trademarks of Meta Platforms, Inc. Relay Port is not endorsed by or affiliated with Meta.