Privacy Policy

3.1 Business Data from Integrations

When you connect third-party platforms, we may access:

E-commerce platforms (Shopify, Centra, etc.):

• Product catalog (names, descriptions, SKUs, prices)

• Store configuration

• Anonymized transaction data (sales, revenue, number of transactions)

Analytics platforms (Google Analytics, Shopify Analytics, etc.):

• Website traffic and user behavior

• Conversion data

• Attribution metrics

Public data collection:

We may also collect publicly available product information from the internet, including product pages, descriptions, and other publicly accessible data.

Data from e-commerce platforms, analytics platforms, and public sources is commercial data, not personal data. We don't collect end-consumer information (customer names, emails, addresses, or payment details). Transaction data is anonymized and aggregated – we never access individual customer purchase details.

Slack is the exception. Data from a connected Slack workspace includes personal data of the workspace's members and is described in section 3.3.

3.2 AI Platform Queries

We send product data to AI platforms (OpenAI, Google Gemini, etc.) to analyze product visibility. Product data is commercial (product names, descriptions) – not personal data.

If you connect Slack, we also send Slack message text, and facts learned from it, to AI platforms. Section 3.3 explains how.

3.3 Slack Data

This section applies when a customer connects a Slack workspace to Glara.

What we collect:

• From each public channel the customer turns on: the message text, the author's Slack ID and display name, the channel, and the time. When a channel is turned on, we collect its messages from the last 90 days, and after that its new messages.

• Messages that mention Glara or are sent to Glara directly, so Glara can reply.

• The names, topics, and purposes of the workspace's public channels, so the customer can choose which channels to turn on.

Glara never reads private channels or channels shared with other organizations.

Why: To learn lasting facts about how the customer's company works ("Memory"), which Glara's agents use when they work for that customer.

Who decides: The customer's company is the controller of its Slack data. Glara processes it on the customer's behalf.

How long we keep it:

• Messages waiting to be processed are kept for up to 24 hours.

• Stored messages are deleted 30 days after Glara processes them.

• Facts in Memory are kept until the customer disconnects Slack.

• When the customer disconnects Slack or uninstalls the Glara app, we delete that workspace's data within 14 business days, backups included.

Who processes it:

• Supabase, Render, and Trigger.dev store and process Slack data in Frankfurt, Germany.

• Message text is sent to OpenAI through Vercel's AI Gateway. Facts in Memory are also sent to Google's Gemini models. These AI providers may process the data outside the EU.

• OpenAI and Google may keep prompts, including Slack message text, for up to 30 days to monitor abuse.

No AI training: We don't use Slack data to train AI models, and our agreements with OpenAI and Google don't allow them to train on it.

During the beta: While the Slack integration is in beta, Glara staff may read stored messages to check and improve how Glara learns from them.

Your rights: A workspace member can ask for access to, or deletion of, their Slack data by writing to [email protected]. On a deletion request, we delete the person's stored and waiting messages and remove the facts in Memory that came from their messages. The request doesn't reach the messages in Slack itself, which belong to the customer, or the AI providers' abuse-monitoring logs, which expire within 30 days.