Snapshot 191524

Document
Privacy policy
URL
https://plugand.ai/privacy.html
Fetched
HTTP status
200
Content type
text/html; charset=UTF-8
Fetch mode
static
Size
100264 bytes
SHA-256 (raw)
fd52695e23238689c139655d226fe60d29dd5958990ef14c4d63c474e25a445e
SHA-256 (normalized text)
cca433bc43af86944102803b14bbfe8d599839ea6bdd33cf40a81505f1c484c7

Normalized text

Scripts and page chrome removed; this is what change detection compares.

1. Introduction
Welcome to Plug and AI ("we," "our," or "us"). We are committed to protecting your privacy and ensuring the security of your data. This Privacy Policy explains how we collect, use, and protect your information when you use our Slack and Microsoft Teams application, the plai.chat web chat, the PLAI Chat apps for iPhone, iPad and Android, and our other services.
Sections 2 and 3 describe the Slack and Microsoft Teams apps. Section 5 describes plai.chat and the PLAI Chat apps; Sections 4, 6 and 7 apply to all of our services.
2. Information We Collect
Slack and Microsoft Teams Data
Messages: We process messages where our bot is mentioned, as well as one-to-one chats and group chats sent to the bot (on Slack, direct messages), to generate AI responses. We do not save message content to our database or file system, with one exception: Ping tasks and reminders on Slack (see below). To answer follow-up questions, recent thread or conversation context is kept in server memory for about 24 hours after its last activity (removed by an hourly cleanup), and is lost when the service restarts.
Ping tasks and reminders (Slack): When you create a task or reminder with Ping, we store the task text (usually a short summary of the request), the assignee and author IDs, the channel, and a link to the source message in our database so the reminder can be delivered and tracked.
User and Team Information: We store your Slack Team ID, User ID, and Team Name (or, on Microsoft Teams, the equivalent tenant, user, and team identifiers) to manage your account and billing.
Files: If you send attachments to the bot (images for vision analysis, PDFs for document analysis), they are processed entirely in-memory and then discarded. We never save your files or images to our servers, database, or persistent storage.
Channel data (with consent): On Microsoft Teams, if a tenant administrator grants the corresponding consent, the bot can list channels and read channel history via the Microsoft Graph API to answer questions with channel context. On Slack, equivalent access is governed by the scopes you approve at installation.
On Microsoft Teams, message routing and delivery are handled through the Microsoft Azure Bot Framework; the bot is hosted on Microsoft Azure.
Usage Data
We record the model, provider, and cost of each billed request, together with your resulting credit balance, to bill your workspace appropriately.
Billing Information
All payments are processed securely by Stripe. We do not store your credit card information. We only retain a record of your credit balance and transaction history.
3. How We Use Your Information
To provide the AI chat service and generate responses.
To maintain conversation context within Slack threads and Microsoft Teams conversations.
To process payments and manage your credit balance.
To improve the performance and reliability of our service.
4. Data Sharing and Third Parties
Zero Data Retention (ZDR)
Our Commitment (Plug and AI)
We keep our handling of your message content to a minimum:
Logs: Our operational logs can include file names, URLs found in messages, search queries the assistant runs, and a short excerpt from the start of AI responses. They are used for troubleshooting, not to keep a record of your conversations.
No Training: We do not use your data, messages, or interactions to train AI models.
No Conversation Database: We do not store your conversation history or message content in our databases, except Ping task text as described above. Recent context is held in server memory for about 24 hours after last activity to answer follow-ups, then discarded.
Third-Party Providers (OpenRouter & Model Providers)
Third-party AI providers (OpenRouter and downstream model providers) may or may not comply with Zero Data Retention policies independently. To help protect your privacy:
ZDR Settings Available (Slack and plai.chat): ZDR is available on Slack and on plai.chat (web and apps), not yet on Microsoft Teams. For selected models that support it, you can enable OpenRouter's Zero Data Retention setting in your preferences on Slack, or with the "No retention" switch on plai.chat. When enabled, this instructs OpenRouter and downstream providers not to store your data or use it for training.
User Control: ZDR is optional and not all models support it. When ZDR is not enabled, your data may be subject to the standard data retention policies of OpenRouter and the underlying model providers.
Exceptions (Slack): Ping task rewriting uses free models, which do not offer ZDR, and image generation does not apply the ZDR filter.
Note: While we pass ZDR settings through when enabled, we cannot guarantee compliance by third-party providers. Please review their respective privacy policies.
We do not sell your data. We share data only with the following necessary service providers:
AI Model Aggregation
We utilize OpenRouter as our exclusive gateway to access various AI models. When you interact with our service, your message content is transmitted to OpenRouter, which routes it to the model provider of your choice.
Multiple LLM Providers
Our service supports 300+ AI models from multiple providers including OpenAI, Anthropic, Google, Meta, Mistral, and others. The specific model used depends on your preference settings or explicit selection.
For detailed information about the following policies, you must consult OpenRouter's documentation and each model provider's respective policies:
LLM model(s) being used: See OpenRouter Models
LLM data tenancy policy: Varies by provider — review each provider's terms
LLM data residency policy: Varies by provider — data may be processed in different regions
LLM retention settings: See OpenRouter ZDR documentation and individual provider policies
Zero Data Retention (ZDR) settings are available for selected models that support them, and ultimate compliance depends on the model provider. Users requiring strict data governance should review provider-specific policies before use.
Important Chain of Responsibility Notice
Our service acts as an interface between your Slack workspace, Microsoft Teams tenant or plai.chat and OpenRouter. By using our service, you acknowledge and agree to the following data processing chain:
Plug and AI (Us): We process your request and securely forward it to OpenRouter.
OpenRouter: Acts as an aggregator and forwarder. Please review OpenRouter's Privacy Policy. We do not control, are not responsible for, and do not account for OpenRouter's privacy practices or data handling.
Model Providers: OpenRouter forwards your request to the specific underlying model provider you select (e.g., OpenAI, Anthropic, Google, etc.). Each of these providers has their own separate privacy policies and data retention practices.
Disclaimer: We expressly disclaim liability for the privacy practices, security standards, or data handling of OpenRouter and any downstream model providers in this chain. You utilize these third-party services at your own risk and are encouraged to review the privacy policies of the specific AI models and aggregators involved.
Request Classification (TypeSafe)
On the plai.chat website we send a sample of messages to TypeSafe (the Jev model), which estimates what a message is asking for — for example whether it wants an image generated, whether answering it needs up-to-date information from the web, or whether a diagram or an interactive page would serve you better. Today we use the result only to measure whether such automatic classification could help us choose tools and models in the future; it does not change how your request is answered.
What is sent: the text of up to your three most recent messages in the current conversation, and a single yes/no flag saying whether an image already exists in that conversation.
What is not sent: the assistant's replies, your attachments (documents and images), your email address, and your account identifier. The classification is a probability, not a stored profile.
Zero Data Retention: if you have ZDR enabled, your messages are not sent to TypeSafe at all.
Their policy: see TypeSafe's legal documentation.
This does not apply to the PLAI Chat apps for iOS and Android, the Slack app or the Microsoft Teams app.
Infrastructure & Payments
Heroku (Salesforce, Inc.): We use Heroku for application hosting.
Cloudflare: We use Cloudflare for database services and CDN.
Stripe: We use Stripe for secure payment processing.
Slack and Microsoft Teams: As a Slack and Microsoft Teams app, all interactions occur within those platforms.
plai.chat and the PLAI Chat apps use further providers (for example Apple, Google and RevenueCat for in-app purchases and notifications); they are listed in Section 5.10.
4b. Embedded Widget on Third-Party Websites
This section applies only to one specific way of using our service: the embeddable PLAI.chat chat widget, which a website owner can add to their own site with a <script> tag. It does not apply to the Slack app, the Microsoft Teams app, or to using plai.chat directly.
When you chat inside a widget embedded on someone else's website, we process, in addition to the message content described above: the address (URL) and title of the page the widget is on, so the assistant can answer questions about it; the origin of the embedding site and your IP address, for abuse prevention and rate limiting; and a short-lived, randomly generated visitor session identifier. Conversations remain stored in your own browser. The widget sets no cookies on the host website.
The operator of the website hosting the widget decides whether to enable it and which features to allow; for that data we act as their processor under our Data Processing Agreement, which also lists the sub-processors involved. A countersigned copy can be requested via our Support page.
Every embedded widget visibly carries a "Powered by plai.chat" label, so you can always tell when a chat on another website is served by us.
5. plai.chat and the PLAI Chat Apps (iOS and Android)
This section applies to plai.chat on the web and to the PLAI Chat apps for iPhone and iPad and for Android, distributed through the App Store and Google Play — "the apps". The apps show plai.chat inside the app, so the same account, credits and rules apply, plus the app features described below. Section 4 (OpenRouter, model providers, Zero Data Retention) applies as well.
5.1 Chats, Attachments and AI Processing
What is sent: to answer you, your message, the earlier messages of the same chat, and any photos, PDFs or other files you attach or share into the app are sent to our server and forwarded to OpenRouter and to the provider of the model that answers (see Section 4). Whether that provider may keep the data depends on its own policy, unless you turn on the "No retention" (Zero Data Retention) switch.
Chats stay on your device: conversations are saved in your browser or in the app's storage on your device, not in our database. The only exception is the next point.
Answers finished in the background (apps, signed in): if the app goes to the background or loses its connection while an answer or image is being generated, our server keeps the finished result in memory for up to 30 minutes so the app can collect it. It is deleted when the app has collected it, after 30 minutes, or when the server restarts.
Read aloud: the speaker button uses the text-to-speech built into your browser or device. We do not receive the text; depending on the voice, your browser or device provider may process it.
Voice dictation: the microphone button uses the speech recognition built into your browser or device (Apple speech recognition on iPhone and iPad, the Android speech recognizer, or your browser's speech service). That provider processes the audio under its own terms; we receive only the resulting text, and only if you send it.
Whiteboards: when an answer you have just received contains a whiteboard or task board, its content is sent automatically to cnvs.app (a whiteboard service from our developer, Coder AI) to create the board. For older or saved answers this happens only when you tap "Create board". Anyone with the board link can view it.
No training: we do not use your chats to train AI models.
5.2 Using plai.chat Without an Account
You can chat with free models without signing in. To keep the free tier fair and stop abuse:
Cloudflare Turnstile checks that a real person is using the service. It evaluates technical signals from your browser or device; after a successful check we set a session cookie that lasts 2 hours.
Daily free-message limits are counted per device (a random identifier in the plai_did cookie, valid for one year), per IP address and, when signed in, per account. We store only hashed (pseudonymised) forms of these identifiers with a daily count — never message content.
For each request without an account we record a hashed session identifier, the model used, its cost and the time, to detect abuse and billing errors.
5.3 Your Account
Sign-in: you sign in with your email address. We email you a sign-in link that expires after 15 minutes; emails requested from the apps also contain a 6-digit code, which we store only in hashed form. To protect accounts against guessing, we keep a record of wrong codes (the email address and the time) for 24 hours. You can also add passkeys: we store the passkey's public key and identifiers, never your fingerprint, face or device PIN.
Account data: email address, account ID, credit balance, total spend, and the times of sign-up and last sign-in.
Credit ledger: each credit transaction — welcome credit, purchase, refund, and every paid request with the model used, its cost and your resulting balance. The ledger contains no message content.
Sessions: a sign-in cookie keeps you signed in. On the website a session ends 30 days after you signed in. In the apps it is extended while you use them: it ends after 30 days without use and in any case no later than 180 days after you signed in. Signing out ends it on that device.
5.4 Purchases
On the web, credits are paid through Stripe. Stripe processes your payment details; we receive the amount, currency and a payment reference, never your card number.
In the apps, credits are sold only as in-app purchases billed by Apple (App Store) or Google (Google Play), which process the payment under their own terms and privacy policies. We never receive your card or payment details.
RevenueCat (RevenueCat, Inc.) validates in-app purchases for us. The app gives RevenueCat your plai.chat account ID; RevenueCat receives the purchase details from Apple or Google (such as the product, transaction identifiers, price, currency and store country) and technical information from the app needed to process the purchase, and tells our server to add the credits. Refunds and reversed purchases are reported the same way, and we then remove the credits that purchase added (up to your remaining balance).
Records: we keep purchase records in our credit ledger for accounting. RevenueCat keeps a customer record (your account ID and purchase history) and Apple and Google keep their own purchase records under their own policies. Deleting your plai.chat account does not automatically delete the RevenueCat record — ask us through our Support page and we will request its deletion.
5.5 Notifications (Apps)
If you allow notifications, the app registers with Apple Push Notification service (iPhone and iPad) or Firebase Cloud Messaging by Google (Android) and sends us the device's push token, its platform and the app's language. We use them only to send you notifications about your own requests and purchases — for example that an answer or image is ready. Notifications never contain the content of your chats, and we send no marketing notifications. You can turn notifications off at any time in your device settings. A token is deleted when you sign out on that device, when you delete your account, or when Apple or Google report it as no longer valid.
5.6 Reporting AI Answers (Apps)
The Report button under an AI answer or generated image sends us the reason you choose, your optional note, the model, the type of answer (text or image), your platform, the time and — if you are signed in — your account ID. If you are signed in and tick "Include this answer in the report", the text of the answer (up to 8,000 characters) is included too; reports sent without an account never contain it, and generated images are never sent. We use reports to review and prevent harmful output. We keep them for 90 days, and we delete the reports linked to your account when you delete it.
5.7 Analytics, Cookies and Device Storage
Website analytics: the plai.chat website uses Google Analytics (Google) to measure visits. It sets cookies and receives your IP address and technical information about your browser. The website also loads a script from coderai.dev (our developer, Coder AI) for the "Report a feature or bug" form and the list of our other apps; what you submit in that form goes to Coder AI. This privacy policy page on plugand.ai also loads the coderai.dev feedback widget script for its "Contact us" link; what you send through it goes to Coder AI. You can block these cookies and scripts in your browser settings.
The apps contain no advertising, no third-party analytics and no tracking, and do not load Google Analytics or the coderai.dev scripts. TypeSafe classification (Section 4) is not used for requests from the apps.
Cookies we set: the sign-in session cookie, the 2-hour session cookie after the Turnstile check, and the plai_did device cookie for daily free limits (Section 5.2). Cloudflare Turnstile may store its own technical data to tell people from bots.
Device storage: your chats, settings and choices (for example your theme and whether you agreed to the AI data notice in the apps) are stored on your device. On Android, the app's data (including saved chats, but not the sign-in session) can be included in your device backup; on iPhone and iPad, app data can be included in your device backups. You control those backups.
We do not show ads and we do not sell your data.
5.8 Deleting Your Account
You can delete your plai.chat account yourself, on the web or in the apps: sign in, open the account menu (the person icon at the top right), choose Delete account and type your email address to confirm. When you do:
your sessions, passkeys, pending sign-in links, push tokens, content reports and records of wrong sign-in codes are deleted;
your email address is replaced by a pseudonym (a keyed hash) that we keep only so the same address does not receive a second welcome credit;
your remaining credit balance is forfeited and cannot be refunded;
an active plai Plus subscription (sold on the plai.chat website) is cancelled with Stripe immediately and is not charged again;
we keep the credit ledger of past purchases and usage for accounting and fraud prevention, linked only to the pseudonymised account, not to your email address; RevenueCat, Apple and Google keep their purchase records as described in Section 5.4;
chats saved on your devices are not affected — delete them in the app or your browser, or uninstall the app.
If you cannot sign in, you can ask us to delete your account through our plai.chat support page.
5.9 Age Requirement
plai.chat and the apps are intended only for people aged 18 or over. We do not knowingly collect data from anyone younger.
5.10 Service Providers for plai.chat and the Apps
OpenRouter and the AI model providers — generating answers and images (Section 4).
Heroku (Salesforce, Inc.) — hosting.
Cloudflare, Inc. — CDN, Turnstile bot protection and the database.
Mailgun (Sinch) — sending sign-in emails.
Papertrail (SolarWinds) — application logs.
Stripe, Inc. — payments on the web.
Apple and Google — app distribution, in-app payments and push notifications (Apple Push Notification service, Firebase Cloud Messaging).
RevenueCat, Inc. — validating in-app purchases.
Google Analytics — website only, not the apps.
TypeSafe — sampled request classification, website only (Section 4).
cnvs.app — only when an answer contains a whiteboard (Section 5.1).
Google (Google Forms) — the contact form on the support page (support and account-deletion requests, including the email address you enter).
Some of these providers process data outside the European Economic Area, mainly in the United States. Such transfers rely on the European Commission's Standard Contractual Clauses or another valid transfer mechanism available to that provider. All connections between your device, our servers and these providers are encrypted in transit (HTTPS/TLS).
5.11 Logs
Our server logs record technical events such as requests, errors, IP addresses, the model used and its cost, and sign-in events, which can include your email address. As described in Section 4, they can also include file names, URLs, web search queries and a short excerpt from the start of an AI answer. We use them for troubleshooting, security and abuse prevention, and keep them for up to 7 days.
6. Data Retention
For the Slack and Microsoft Teams apps, we retain account and transaction records (billing data) as long as your account is active.
For plai.chat and the apps:
Account data — until you delete your account (Section 5.8).
Credit ledger and purchase records — kept for accounting, tax and fraud prevention, also after account deletion (pseudonymised).
Sessions — 30 days on the website; in the apps 30 days without use, at most 180 days; sign-in links 15 minutes; records of wrong sign-in codes 24 hours; passkeys until you remove them or delete your account.
Push tokens — until you sign out on that device, delete your account, or the token stops being valid.
Background answers — up to 30 minutes, in memory only.
Content reports — 90 days (Section 5.6); deleted earlier with your account.
Hashed free-limit counters and anonymous usage records — as long as needed for abuse prevention and usage statistics; they contain no message content.
Logs — up to 7 days.
Chats — on your device, until you delete them.
7. Your Rights
You can request the deletion of your data by contacting us. Slack workspace or Microsoft Teams tenant administrators can uninstall the app at any time; uninstalling stops the bot from receiving further messages. On Microsoft Teams, any Microsoft Entra admin consent previously granted (for example Microsoft Graph permissions) can be revoked separately by the tenant administrator in the Microsoft Entra admin center. Uninstalling does not by itself delete stored account and billing records — to have that data deleted, please request deletion via our Support page.
plai.chat accounts can be deleted directly in the web app or the apps (Section 5.8).
Where data protection law such as the EU General Data Protection Regulation applies, you also have the right to access your data, to have it corrected, to receive it in a portable format, to restrict or object to its processing, and to lodge a complaint with a data protection authority — in Poland, the President of the Personal Data Protection Office (UODO). To exercise these rights, contact us through our Support page.
8. Contact Us
If you have any questions about this Privacy Policy, please visit our Support page.
Coder AI is a brand of sp. z o.o., Wadowicka 7, 30-347 Kraków, Poland · EU VAT:
· Contact us