Privacy Policy

Last Updated: October 30, 2026

1. Introduction

SynchroVerse ("we," "our," or "us") is committed to protecting your privacy. This Privacy Policy explains how we collect, use, disclose, and safeguard your information when you use our AI companion platform and related services.

2. Information We Collect

2.1 Information You Provide

  • Account Information: Name, email address, password, profile picture
  • Date of Birth: Asked during account creation to confirm you are 18 or older. We use it only for that check and do not keep your birthdate.
  • Payment Information: Billing and payment-method details, processed by Stripe.
  • Avatar Customization: Preferences, personality settings, voice selections
  • Conversation Data: Text and voice interactions with AI companions
  • Support Communications: Messages you send to our support team

2.2 Waitlist and Former $ARIA Data

  • Waitlist Sign-ups: If you join a waitlist on our website, we collect your email address, which waitlist you joined, the referring page, and campaign tags, and we use that information to manage the waitlist and contact you about it. We also record the IP address and browser information (user agent) your sign-up came from.
  • ARIA Waitlist and $ARIA Payments: The ARIA waitlist is closed, and $ARIA token payments are not currently available. If you provided an email address, X (Twitter) account, or Solana wallet address for those programs, we may still keep it. You can ask us to delete it using the contact details below.

2.3 Information Collected Automatically

  • Device Information: Browser type, operating system, device identifiers
  • Usage Data: Features used, session duration, interaction patterns
  • Log Data: IP address, access times, pages viewed
  • Cookies and Browser Storage: one login cookie that keeps you logged in, plus browser storage (local and session storage) for your login, your cookie choice and your settings, and, only if you allow Analytics, a telemetry session ID and the Google Analytics cookies (_ga); see our Cookie Policy for the full list.
  • Website Analytics (optional): only if you turn on Analytics in the cookie settings, Google Analytics counts visits and the pages viewed on synchroverse.ai. It receives page addresses without login codes or other private link details, a general page title on logged-in pages, the referring site, browser and device details and an approximate location (Google sees your IP address). We do not send it your account ID, email, name, agent names or conversations, and its advertising features are turned off. Turning Analytics off stops it.

3. How We Use Your Information

We use the collected information to:

  • Provide, maintain, and improve our Service
  • Process transactions and send related information
  • Enable AI companion memory and personalization features
  • Send technical notices, updates, and support messages
  • Respond to your comments, questions, and requests
  • Monitor and analyze trends, usage, and activities
  • Detect, investigate, and prevent fraudulent or unauthorized activities
  • Comply with legal obligations

4. AI and Conversation Data

Important: Your conversations with AI companions are processed to provide the Service. Here's how we handle this data:

  • Memory Feature: Conversation summaries may be stored to enable cross-session memory
  • Voice Data: Audio is processed to recognize speech and generate replies. If you use voice cloning, your uploaded sample and voice profile are also processed to create and use the voice. The services involved depend on the feature and session configuration; see Section 5.1.
  • Third-Party AI: The providers and information they receive are listed in Section 5.1, including processing for conversations, memory, searchable knowledge and optional voice or image features.
  • No Training: We do not use your personal conversations to train our AI models without explicit consent
  • Game History: Saving game conversation words in account History is off by default. You can opt in under Settings → Privacy for new sessions, then turn it off at any time. Available words are kept for up to 90 days and can be deleted in Conversation history. Some game sessions may provide no conversation words or only part of the conversation; enabling this setting does not recover earlier conversations.
  • Session Logs: Privacy mode is on by default. With it on, session logs are designed to keep technical details such as timing and events while leaving out the words you and your avatar say. If you turn it off in Settings, new session logs can also keep those words. You can view, export and delete your own session logs in Settings.
  • Conversation History: In live avatar sessions, your messages and the companion's replies are saved word for word in your conversation history, so the conversation can be shown again. If your account belongs to an organization, the organization's owners and admins can see a list of members' conversations, with their titles and the first 100 characters of each conversation's latest message. If you save a transcript to an agent's knowledge, that copy belongs to the agent: people in the agent's organization can see it, the agent can use it in its answers, and it is kept until it is deleted from the agent's knowledge or the agent is deleted.
  • Crisis Referrals: In live avatar sessions, a message that starts a crisis referral can be saved in your conversation history like any other message. In text chat through our API we do not store its words. We keep an anonymous daily count of referrals from text chat through our API, and are designed to keep it for live avatar sessions too, with no account, name, or message text. In session logs the message is designed to appear only as "[crisis referral given]". See our Safety page.
  • Organization-Supplied Engine Credentials: Where this feature is available, organizations can supply credentials for ElevenLabs, xAI, or OpenAI instead of using the included SynchroVerse AI. We store those credentials encrypted for the organization, never send stored values back to the browser, and use them only server-to-server to run the selected provider. Conversations in those sessions are processed by that provider.

4.1 Camera and screen sharing

You can choose to share your camera or a screen, window, or browser tab during a live session. Sharing starts only when you turn it on, and you can stop it at any time.

Your video is transmitted to the live session. Our AI analyzes still images from it, resized when needed, to respond to what you show and to visible expressions and body language. The frequency varies with the session and AI model. These are automated interpretations and can be wrong.

With the included SynchroVerse AI, image analysis does not run on your device. Depending on how the companion is configured, it runs on the server hosting your companion, provided by our cloud GPU provider, or the images are sent for analysis to a cloud AI service we use, such as Google's. Where available, sessions using an AI provider connected by you or your organization send audio to that provider; shared video may also be accessible to that connection, depending on the integration.

Shared images are used to provide the live session. Temporary images and recent text descriptions may be held temporarily to provide the session; how they are handled depends on how the companion is set up, as described above. The companion's replies about what you share are part of the conversation and can appear in saved conversation data or, if you enable them, logs containing conversation text, as described in Section 4. A saved memory may note that a visual observation occurred.

Share only the screen, window, or tab you intend to show. It may contain private information about you or other people; avoid sharing sensitive information or others' information without permission. Screen sharing sends the picture only; sound playing on your screen is not shared with your companion.

We do not use these images to identify you, create a face print, or train AI models. Stopping sharing stops new capture; it does not delete information already saved in conversation data. See Sections 7 and 8 for retention, access, and deletion, or contact privacy@synchroverse.ai.

4.2 Optional cloud AI

SynchroVerse AI runs on the server hosting your companion, not necessarily your device.

Google receives conversation text and context to answer when you turn on optional cloud AI and use your Google AI key.

Speech recognition and voice stay on the session host unless you choose a voice or speech service from ElevenLabs, Cartesia, xAI, OpenAI or Deepgram (see the service list).

Stored provider keys are never sent back to the browser. When you use your own AI key, we pass it securely to the conversation so your provider can answer.

5. Data Sharing and Disclosure

We may share your information with:

  • Service Providers: Third parties that perform services on our behalf (payment processing, hosting, analytics)
  • AI Partners: Third-party AI services necessary to provide our features
  • Cloud GPU streaming provider (avatar streaming): When you watch or talk to an avatar, our cloud GPU streaming provider processes the avatar audio/video stream and related session data, including incoming camera and screen sharing for image analysis; depending on the feature used, this may include conversation text or microphone audio, your account identifier, generated stream and session identifiers, our project identifier with that provider, the page URL, device, operating-system and browser information, WebRTC connection data (including network addresses and stream-quality statistics), and approximate location data associated with your connection (country, state or region, city, postal code, and latitude/longitude).
  • Legal Requirements: When required by law or to protect our rights
  • Business Transfers: In connection with any merger, sale, or acquisition

We will never sell, rent, or share your personal information with third parties for marketing purposes.

5.1 Services that process information

These are the outside services used to operate the platform and its available features. Which ones receive your information depends on the feature, session configuration and connected accounts you use. A service listed for an optional feature does not receive every conversation. Some services are connected by you or your organization and handle information under their own terms and privacy policies.

Sentry
Where configured, receives error reports and request diagnostics to help us find and fix service failures. These can include account identifiers, URL paths, timing, browser and device information. Optional browser monitoring runs only after Analytics consent and can include a masked recording of page interactions. Text and input masking and privacy filters reduce the data sent. Server error monitoring is separate from optional browser Analytics. Browser replay masks text and blocks media; direct browser connections also expose your IP address.
Amazon Web Services (AWS)
Hosts our website, APIs, databases, file storage and live-session infrastructure. It processes account details, conversations, saved memories, uploaded files and voice samples, and connection and service logs needed to run those services. Amazon Cognito handles account authentication, including email addresses and login details. Amazon Simple Email Service (SES) receives recipients and email contents to send verification codes, account notices, invitations and other service emails.
Google
Receives conversation text and context for AI responses and memory extraction where those features use Google. It also receives knowledge documents, saved-memory text and search queries to create embeddings (numeric representations that make text searchable). Depending on the session, Google may receive microphone audio for speech recognition or voice interaction, text to generate speech, and still images from camera or screen sharing you turn on. Support text and attachment links may be sent to Google for security checks. Google also handles Google login, search requests and font downloads where used; login shares basic profile details with us, and font requests expose your IP address and browser information to Google. Only if you turn on Analytics in the cookie settings, Google Analytics receives the addresses of the pages you view on synchroverse.ai (without login codes or other private link details), page titles (a general title on logged-in pages), the referring site, browser and device details and an approximate location, to count visits; it sets _ga cookies and sees your IP address. We do not send it your account ID, email, name, agent names or conversations, and its advertising features are turned off.
OpenAI
Receives conversation text, instructions and relevant context when an OpenAI-backed response or connected voice session is used. Voice sessions can send microphone audio, and supported sharing integrations can make shared images available. OpenAI also receives document, memory or query text when it is the configured embedding provider, to make that text searchable.
xAI
Receives conversation context and microphone audio when your organization connects an xAI voice account. Search, research and image-generation features that use xAI send the relevant query, topic or image description so it can return an answer or generated image.
ElevenLabs
Receives text and voice settings to generate speech when selected. Connected conversational sessions also send microphone audio and conversation context. If you create a voice clone with ElevenLabs, it receives the uploaded voice sample and the voice name or description needed to create and manage the clone.
Cartesia
Receives the text to speak, voice identifier, language and voice settings when selected for speech. Voice cloning sends your uploaded audio sample and voice details to create and manage a reusable voice.
Deepgram
Receives microphone audio to turn speech into text when selected for speech recognition, including directly from your browser in supported sessions. Direct browser connections also expose your IP address. For automatic or retried voice-reference transcription, our speech-recognition processor receives the selected passage from your uploaded or recorded audio and returns text after you confirm recording rights. You can avoid automatic transcription by typing the transcript before confirming recording rights; retrying transcription sends the passage again. We exclude these reference-transcription requests from model improvement. The processor documents that the audio and transcript are kept only as needed to process the request; request metadata and usage logs can remain. This does not delete recordings or transcripts you save with a voice on our servers. When selected for speech generation, it receives the text to speak and voice settings.
RunPod
Provides computing for voice generation and voice cloning. When you upload a recording to create a cloned voice, RunPod receives that recording, any supplied transcript and voice settings, and returns a voice profile and a short preview. RunPod's documentation says results of asynchronous jobs remain available for 30 minutes after the job completes. That is RunPod's published default for job results, not a retention period we set. We also set a 10-minute job time-to-live from submission, after which RunPod's docs say the job is deleted. We have not independently verified how long RunPod keeps the original uploaded recording on its systems. Our clone worker returns only a voice profile and a short preview and does not keep a separate copy of the recording after the job ends. We still keep the uploaded recording, the profile, and the preview on our own servers until you delete the voice. Other jobs can send the text to speak and reference audio. This includes features using AI software that we operate on rented servers.
Cloud GPU streaming provider
Hosts and streams live avatars. It processes avatar audio and video, session identifiers, and the conversation text, microphone audio or shared camera and screen content used by the session. It also receives account and project identifiers, the page URL, browser and device details, network addresses, connection-quality measurements and approximate connection location, as described in Section 5.
Stripe
Processes payments, subscriptions and billing support. It receives customer and billing contact details, payment-method information, the plan or purchase, and transaction information. Checkout and billing pages also collect the browser and connection information needed to operate and protect payments.
Cloudflare
Turnstile checks help protect forms and login from automated abuse. Where the check loads, Cloudflare receives IP addresses, browser and device signals, the site requesting the check and verification results. A Cloudflare Worker also relays live session audio for some avatar conversations; that path receives the session audio and your IP address.
HubSpot
Receives contact and demo-request details (name, email, company and message) as our CRM, so we can respond to and manage the request.
Slack
Receives the same contact and demo-request text, including the page it was sent from, as an internal notification to our team.
Asana
Receives support requests, your name, email address and account ID, and the title, agent, message count and date of your 10 most recent conversations. A title can contain the first words of a message. Attachments are uploaded to Amazon S3 and added to the Asana request. It also receives demo-request details (name, email, company and message) so our team can follow up. This helps our team investigate and respond.
n8n Cloud
Runs connected automations where enabled. It receives the instructions, action inputs and results needed for the workflow, which can include message text, recipients, document content and connected-account information. The services that the workflow calls also receive the data needed for their actions.
Vercel
Hosts a service used for connected tools. This is not our website host. Requests routed through it include the tool name, inputs such as search queries or URLs, and any connected-service credentials needed for that request, so the tool can call the selected service and return its result.
WalletConnect
Where a previously linked Solana wallet reconnects in the app, WalletConnect can receive your IP address and public wallet address. Token payments are currently turned off; the wallet connection layer can still run.
Solana
The public Solana network RPC used by the wallet layer can receive your IP address and public wallet address if a wallet reconnects.

We operate our own memory service and a live audio/video relay on hosted infrastructure. Some avatar sessions also send live audio through a Cloudflare Worker, named above. Using software we operate does not mean that all processing happens on your device or that no outside service receives data. We are trialing in-house embeddings, but Google remains in use for embeddings and other AI processing; this trial does not remove Google from the disclosures above.

5.2 Services used by connected features

When an enabled tool or integration runs, it sends the inputs needed for that request. Some features require your own connected account or key. These services receive data for the feature being used, rather than automatically receiving your whole conversation:

Search and reading
Google, DuckDuckGo, Brave Search, Serper, SerpApi, Tavily and Firecrawl receive queries, research topics or page URLs when the corresponding search or reading feature is used. A website you ask us to read also receives the page request. Queries can contain personal details if you or the companion include them.
Weather, reference, news and financial data
Open-Meteo and OpenWeatherMap receive the place or coordinates requested for weather. Wikimedia receives encyclopedia search terms. NewsAPI receives news search terms. CoinGecko, DexScreener, Etherscan, Alpha Vantage and Yahoo Finance receive the selected asset, trading pair, wallet or contract address, ticker, or other lookup parameters when those features are used. Where AI market analysis is enabled, Anthropic receives the selected asset's public market data to produce that analysis.
Trivia and jokes
Open Trivia Database receives the requested category, difficulty and question type. JokeAPI receives the requested joke category and safety filters. These requests retrieve content for the companion to use.
Streaming and connected channels
If you connect or broadcast to Twitch, YouTube (Google), Kick, Facebook (Meta), X, Rumble, or Restream, the chosen service receives the account authorization and channel details, messages, moderation actions, or broadcast audio and video required by that feature. Other streaming destinations you configure receive the broadcast you send to them. Content you publish can be visible to that service's audience.
Your other connected services
When you enable an email or workspace action, the service you connect receives the recipients, message, document or other action inputs needed to carry it out. These may include Gmail (Google), Outlook (Microsoft), another email provider, Slack, Discord, Notion, Asana, Jira (Atlassian), HubSpot or Salesforce. Custom connections send the requested data to the service you or your organization choose.

Voice cloning

What we save for voice cloning

If you create a cloned voice, we keep the uploaded recording, supplied transcript, resulting voice profile and saved preview on our servers, including cloud infrastructure we operate or rent. An unsaved in-session clone is kept as a private draft for 24 hours, then those four content items are deleted automatically. We also keep the voice's name, settings, account and organisation association, processing status and a record of the consent text version and time you accepted it. We use these data to create and provide the voice and handle security, consent and rights issues. If you save and enable organisation sharing, other authorised members may be able to use the voice. Creating a clone does not itself authorise publication to a public voice marketplace.

Where voice data is processed and by whom

Creating a clone can involve sending the recording, any supplied transcript and voice settings to our cloning processor. When you use it, the voice profile and related data are processed by the infrastructure and providers supplying the selected voice or avatar feature. The route depends on the feature and configuration.

AWS hosts our storage and other service infrastructure. RunPod provides voice-cloning and voice-generation computing and receives the data described in its entry in Section 5.1. Our cloud GPU streaming provider hosts and streams avatar sessions and processes their audio, video and related session information. Cloudflare may relay live session audio. If a selected feature uses ElevenLabs, Cartesia or another named service, its corresponding entry in Section 5.1 also applies. These companies are named on our legal pages; the full service list explains the information each receives and why. Session processing and creating the clone are different steps; RunPod enrollment is not described as fallback-only.

Voice retention and deletion

We keep an unsaved in-session voice as a private draft for 24 hours. If it is not saved by its deadline, its recording, transcript, profile and preview are deleted automatically; the deletion audit keeps non-content facts and hashes only. Saving makes it a normal durable voice. We keep a saved voice’s recording, transcript, profile and preview while the voice is kept in your account, subject to the permission, withdrawal and retention rules below. On successful voice deletion, these are removed from our active database. If you delete your account, access to your account is disabled when the deletion request is accepted and permanent deletion is scheduled after a 30-day grace period; voices you own are removed with your account. Shared voices may remain available to authorised organisation members during that grace period unless a separate removal, withdrawal or complaint requires use to stop.

To remove a voice sooner, use the voice-deletion control before closing your account or contact privacy@synchroverse.ai. Where processing relies on consent, the speaker may withdraw it at any time; withdrawal does not affect processing that was lawful beforehand. We will stop the processing that relies on that consent and handle deletion within the applicable legal requirements. Withdrawal does not require closing an account. A valid removal or withdrawal must prevent new use of the affected clone even if erasing residual copies takes longer. Existing audio already lawfully received by others cannot be recalled by deleting the clone. We may retain limited records on a separate lawful basis to meet legal duties or resolve claims; that does not permit continued generation. Processor and backup copies must be handled under the applicable retention and destruction rules, rather than kept indefinitely because a deletion failed.

We will assess consent-withdrawal notices and voice-rights complaints without undue delay. Once we identify the affected clone from a withdrawal request by the speaker or their authorised representative, or from a complaint giving reasonable grounds to suspect unauthorised use, we will promptly pause new generation and use of that clone while we review the matter. We will stop affected in-flight generation and playback within systems we control, including processing operated for us by our service providers. We cannot recall audio already received by others or stop playback outside those systems. Any verification must be proportionate and must not delay urgent protective action or compliance with applicable law. Residual storage does not authorise continued use. We may restore use only after establishing the permission and lawful basis needed for it.

Erasing residual copies may take longer than stopping use. We do not promise instant erasure of all processor or backup copies. This does not excuse any legally required deletion deadline.

Consent records, verification and complaints

We record which clone consent text you accepted and when. This is your attestation; it is not proof that we have independently verified the speaker. If verification or a rights complaint requires more information, we will explain what we need and how to provide it. The same notice and authorisation requirements apply whether covered sensitive or biometric processing occurs during cloning, use or verification; they are not limited to verification checks. Where a separate lawful basis permits or requires it, we may retain limited consent, complaint and enforcement records after a voice or account is deleted to establish what was authorised, investigate misuse or meet legal obligations, only for the period that basis permits or requires. Retaining evidence does not permit new generation or other continued use of the clone. Routine deletion audit records contain identifiers, status and consent metadata rather than the voice recording or profile. Report unauthorised voice use through Support or legal@synchroverse.ai; use privacy@synchroverse.ai for privacy-rights requests.

voice-cloning Terms · Acceptable Use rules · Report unauthorised voice use · legal@synchroverse.ai · privacy@synchroverse.ai

6. Data Security

We implement appropriate technical and organizational security measures to protect your data, including encryption in transit and at rest, secure authentication, regular security assessments, and access controls. However, no method of transmission over the Internet is 100% secure, and we cannot guarantee absolute security.

7. Data Retention

We retain your information for as long as your account is active or as needed to provide services, except that temporary images and descriptions used during a live session are handled as described in Section 4.1, and except that some data, such as older conversation history and avatar session logs, may be deleted sooner on a set schedule. You may request deletion of your data at any time, and you can delete your avatar session logs yourself in Settings. When your account is permanently deleted, your conversation history is deleted with it (a transcript saved to an agent's knowledge stays with the agent, unless you created that agent: agents you created are deleted with your account, together with their knowledge; see Section 4). Some information may be retained as required by law or for legitimate business purposes (e.g., fraud prevention, legal compliance).

8. Your Rights and Choices

Depending on your location, you may have the right to:

  • Access, correct, or delete your personal information
  • Object to or restrict certain processing
  • Data portability (receive your data in a structured format)
  • Withdraw consent at any time
  • Opt-out of marketing communications
  • Request deletion of conversation history

To exercise these rights, contact us at privacy@synchroverse.ai

9. International Data Transfers

Your information may be transferred to and processed in countries other than your own. We ensure appropriate safeguards are in place to protect your data in accordance with applicable laws, including standard contractual clauses where required.

10. Children's Privacy

The Service is for people age 18 and older. We ask for your date of birth when you create an account and do not create an account if the age check shows that you are under 18. We do not keep the birthdate used for this check. If we learn that we have kept personal information from someone under 18, we will delete it promptly.

11. California Privacy Rights (CCPA)

California residents have additional rights under the CCPA, including the right to know what personal information we collect, the right to delete, and the right to opt out of sale or sharing. We honor Global Privacy Control signals by turning off non-essential categories.

12. European Privacy Rights (GDPR)

If you are in the European Economic Area (EEA), you have rights under the GDPR including access, rectification, erasure, restriction, portability, and the right to lodge a complaint with a supervisory authority. Our legal bases for processing include consent, contract performance, and legitimate interests.

13. Changes to This Policy

We may update this Privacy Policy from time to time. We will notify you of material changes by email or through the Service. Your continued use after changes constitutes acceptance of the updated policy.

14. Contact Us

If you have questions about this Privacy Policy or our data practices, please contact us: