Your discussions, saved to your account. HuddleSpeak saves transcripts
to your account for continued discussion. You can delete your saved content or account from the app.
We do not sell your personal information or use it for advertising.
This shared policy applies to HuddleSpeak and the branded portals it provides.
You are viewing the policy for Light.
It explains what is saved, how voice and AI features work, and the choices you have.
Who we are
HuddleSpeak (“we”, “us”) provides the platform for this app, website, and its branded portals. You can reach us through our
support page.
What your account holds
Account details: your sign-in information and the name and email supplied by your
sign-in provider, such as Google or Apple. You may also provide an email for recaps.
Discussions: transcripts, typed messages, shared links, and files you upload or
create with the panel.
Memory: short summaries that help the panel continue earlier discussions within your account.
Usage and subscriptions: session length, plan, billing period, subscription status,
and payment-provider identifiers. We do not receive your full payment-card details.
We use this information to provide responses, continue discussions, deliver requested recaps,
manage your plan, and operate, secure, troubleshoot, and improve the service.
Voice and audio
When you use the microphone during a session, your speech is sent for transcription so the panel can
respond. You can use typed input instead. HuddleSpeak does not save microphone audio as a session
recording unless you choose to record where that feature is available. Provider processing is described below.
When voice analysis is enabled, microphone audio may also be sent to a third-party voice-analysis provider to estimate
emotion, vocal style, pitch, accent, and age range. These estimates may appear in live chat and guide
the current response; they are not added to saved transcripts, recaps, or memory.
How providers help deliver the service
The providers used depend on the features you use:
AI services: Third-party AI providers may process your messages and relevant
discussion history, memory, images, or files to generate responses.
Speech and voice services: We use third-party providers to transcribe your speech
and generate spoken responses. Transcription providers receive microphone audio; voice-generation
providers receive the text to be spoken, which may include information from your discussion.
Search and email: providers process search requests or deliver recaps when those features are used.
Payment providers: Apple processes App Store purchases and Stripe processes website
subscriptions. They send us subscription status and identifiers to manage your plan across devices.
Providers’ retention and any use of content for improvement or model training depend on their terms
and the settings for the service we use. Deleting content in HuddleSpeak does not automatically delete
copies held by providers, including generated responses or audio. Contact
support for help with provider deletion options.
DeepDiveLM Reader browser extension
DeepDiveLM Reader is an optional extension for Chrome and Microsoft Edge. It works only with DeepDiveLM, one of
the portals HuddleSpeak provides, and lets DeepDiveLM read aloud and discuss the page you have open in a browser tab
you choose, such as a book or an article. This section applies only if you install it.
What it handles: the extension reads only the tab you choose, either from its list of your
open tabs or by clicking its button in your browser toolbar while on that tab. To show that list and read the tab
you pick from it, it asks your browser, once, for access to all websites; your browser shows that request and you
can decline it and use the toolbar button instead. The list of tab titles stays inside the extension’s own window
and is never sent anywhere. The extension reads selectable text visible in the tab you chose, one page at a
time, and hands it to your DeepDiveLM tab. When usable selectable text is unavailable, it takes a picture of
the visible page instead. A question about a picture can also request a screenshot of the same page.
It highlights the passage being read in the source tab and turns the page when you or DeepDiveLM ask. It does
not read any other tab’s content, sign-in details, cookies, or browsing history, and it sends nothing anywhere
except to your DeepDiveLM tab. When DeepDiveLM asks, it asks your browser to check the extension store for a
newer version of DeepDiveLM Reader, as your browser does on its own every few hours; that check is your
browser’s and carries nothing from your tabs.
How the page is used: selectable text skips image recognition. When a picture is needed,
your DeepDiveLM tab sends the picture to our server over a secure
connection to read its text. For languages written in Arabic script, or a page our own text recognition cannot
read, the picture is sent to Google’s Gemini service to read the text instead. The text is then used to read the
page aloud (it is sent to our voice provider, and for some languages first to an AI provider that prepares it
for the voice) and, when you ask about the page, to answer you (it is sent with your question to our AI
providers).
What is kept: neither the picture nor the page text is stored by the extension or on our
server. Each new page replaces the last. Your discussion about the book is saved to your account like any other
discussion, and you can delete it from the app.
The use of information received from the DeepDiveLM Reader extension adheres to the Chrome Web Store User Data
Policy, including the Limited Use requirements. It is used only to read the page aloud and to discuss it with you,
it is shared only with the providers described above for that purpose, it is never used for advertising, and it is
not sold.
Access and sharing
Saved discussions, files, projects, and memory are stored on our server under your account and are
not publicly visible. You sign in to access them. The service operator has administrative access for
operation, support, and security. We take reasonable measures to protect your information, although
no online service can guarantee complete security.
Credits shared between HuddleSpeak and DeepDiveLM. When the same Apple ID, or the same
email address verified by Google or Apple, signs in to both HuddleSpeak and DeepDiveLM, the two accounts share
their paid credits and top-ups. Free credits stay in the app that gives them. Only credits are shared:
discussions, files, projects, and memory stay in each app. Each app shows the other app's plan, credits left, and
renewal date so the shared balance adds up. To stop sharing, contact support.
If you send a recap or share an exported file, the recipient receives a copy. On MA Dimensions,
transcripts are also shared with the portal operator for follow-up, as
described in the notice shown before you join.
Your choices
Manage saved content: clear conversation history and saved files in the app.
Delete your account: use the Account menu or contact support for help. Deletion
removes saved memory, files, and projects from our active storage. Limited account-status and
usage records may remain; provider-held copies follow the terms described above.
Choose text: decline microphone access and continue with typed messages.
Children
HuddleSpeak is not directed to children and is not intended for use by anyone under 13 (or the minimum age
required in your region). We do not knowingly collect data from children.
Changes to this policy
We may update this policy as the app evolves. Material changes will be reflected by updating the “Last
updated” date above.