Sub-Processors

Cloud hosting, edge computing, serverless functions (Workers), and key-value data storage (KV). Processes configuration data, OAuth tokens (encrypted), session data, message routing metadata retained briefly for thread synchronization (message and channel identifiers, timestamps, and a non-reversible one-way fingerprint, never the message text or sender name, retained for 7 days per the Privacy Policy), transient user profile data (display names and avatar images cached for up to 24 hours, auto-expiring, so senders and mentioned users appear by name on the receiving platform), per-channel member lists for bridged mentions (display names, addressable names and user IDs, replaced after each complete refresh and deleted with the channel mapping), and usage metrics. Workers AI: the text of a question typed into the docs assistant on our help pages is processed to find matching help-article passages; SlackBridge does not store it.

Microsoft Teams API integration for message relay and OAuth authentication. Processes Teams tenant information, channel data, user identifiers, and message content (relayed in real time and not stored at rest; only message and channel identifiers, timestamps, and a non-reversible one-way fingerprint are retained for 7 days per our Privacy Policy, never the message text or sender name).

Slack Web API integration for message relay, channel management, and OAuth authentication. Processes Slack workspace and channel identifiers, user identifiers, and message content (relayed in real time and not stored at rest; only message and channel identifiers, timestamps, and a non-reversible one-way fingerprint are retained for 7 days per our Privacy Policy, never the message text or sender name). Slack also hosts our own team workspace, where website live chats (the name and email address a visitor gives, the page and the chat messages) are shown to our support team; see the Privacy Policy for how long they are kept.

Payment processing and subscription management. Processes billing email, payment method details, and subscription status.

Transactional email delivery for account notifications and invitations. Processes email addresses and email content.

Google Workspace for business email, document collaboration, and internal communications. Processes customer correspondence and internal operational data.

Error monitoring for the SlackBridge API. When an error occurs, processes the error details, the request method and path, organization and user identifiers, and the most recent operational log lines. The incoming request's body, headers and query string are never attached, outgoing request URLs are recorded without their query strings, and the log lines never contain the text of bridged messages or the names of people in bridged conversations; log lines about account events can include the account holder's email address. Error reports are retained for up to 90 days.

Product analytics for understanding feature usage, user journeys, and service improvement. Processes user identifiers, organization details, plan information, and in-app behavior, including, for visitors who have accepted analytics (or by default outside the UK, EU/EEA and Switzerland), session recordings of the website and the signed-in dashboard (typed input always masked; outside public website, help and legal pages all on-screen text is masked and images and labelled elements are blocked; retained for up to 30 days) and heatmap data (click positions, scroll depth and sampled mouse movement per page address, with no screen content). Visitors in the UK, EU/EEA and Switzerland are asked first.

Answer checking for the docs assistant on our help pages, when switched on. Receives the text of a question typed into the assistant and the public help-article passages it is compared with, and returns whether each passage answers the question. Receives the question exactly as typed (so anything a visitor includes in it); SlackBridge never sends it contact-form email addresses, IP addresses, account data or bridged message content. Does not use the data to train models (per its published policy).