1. Information we collect
We receive information directly from you through registration, profiles, event participation, feedback and support. We also collect technical information when you use the service and receive login information from providers you choose to connect.
- Account information: name, username, email, authentication identifiers and a hashed password if you use password login. Google or LinkedIn sign-in supplies the information described in its permission screen.
- Profile and event information: biography, photo, header image, professional interests, contact links, visibility settings, your private Agent Me text, event membership and answers, and introduction credits.
- Generated information: suggested matches, introductions, drafts and related inferences produced from profile and event information.
- Communications: feedback, support requests and correspondence with us. Feedback sent with the Feedback button includes the address of the page you were on, your browser type and window size, and a screenshot of that page when you leave it attached.
- Technical information: IP address, basic request and security logs, session information and browser preferences.
- Purchase information, if you buy a paid service: billing contact details, invoice information, selected plan, transaction references and payment status. Any payment provider and the information it receives will be identified at checkout.
Information needed to create an account or deliver a requested feature is required for that purpose; other profile information is optional. Operating profile agents, assessing connections and generating introductions require relevant profile information to be processed, including through the providers identified as actual recipients in section 5. If information or permissions lawfully required for a feature are missing, we may be unable to provide that feature. This does not make unrelated provider training or other optional processing a condition of access where prohibited by applicable law. Do not submit unnecessary sensitive information or private information about others.
2. How and why we use information
We use account, profile and event information to provide the service you request: authenticate you, manage participation, generate introductions, share relevant information and send service notifications. We rely on contractual necessity for these activities.
For paid services, we use purchase information to process payments, manage subscriptions and provide billing support under our contract with you, and retain required accounting records to meet legal obligations.
We use limited technical information, feedback and support records to diagnose problems, improve functionality, prevent abuse and protect the service. We rely on our legitimate interests in operating and improving a secure service, balanced against your rights.
We also process information to comply with legal obligations and, where necessary, protect or defend legal rights. Where consent is required for an optional use, we obtain it separately and allow you to withdraw it. We do not treat account creation as permission for unrelated marketing.
3. AI and introductions
AI uses relevant profile information, event answers and event context to suggest connections and write introductions. Free-text answers may contain identifying information you include. Generated information can be inaccurate; contact us to report a problem or request correction.
Depending on the feature, we use open-weight or proprietary models through a hosting provider’s API or our own infrastructure. Where an external provider runs the model, relevant inputs are sent to that provider and generated responses are returned to us. Using a model does not necessarily mean that its developer receives your information; this depends on who hosts it and the processing arrangement.
What goes to an AI provider, by feature: for matching and introductions, the answers each participant wrote for the event (headline, what they offer, what they seek), the event's name, purpose and briefs, and the organizer's own instructions; for the example introduction, the event's name and purpose; for drafting event answers from Agent Me, the Agent Me text and the participant's name; for updating Agent Me from event answers, the Agent Me text and the answers; for the one-time header image, the Agent Me text. Email addresses and contact links are not sent. Free text can contain anything you choose to write in it.
We do not train models ourselves. Your Agent Me text may be updated from your own event answers by a model, for you; you can turn that off in your account. Introductions are stored with the name of the model that wrote them.
We limit the information sent to what the feature needs. External AI and API services differ in their handling of inputs, outputs and related technical information. We do not guarantee zero retention, the absence of logging or human review, or exclusion from model training, evaluation or service improvement across all providers. You should not assume that information is processed only temporarily or automatically excluded from those uses merely because it is sent through an API.
These statements describe limits on the assurances we give, not a finding that a particular provider retains information or uses it for training. Known provider practices and applicable restrictions are described in section 5. Where a practice has not been verified, we identify that uncertainty rather than present it as a confirmed practice or protection. This notice does not itself authorise a provider to use information for unrelated purposes.
Introductions are suggestions for you to consider. We do not use them to make solely automated decisions with legal or similarly significant effects about you.
4. Who receives information
Participants and organisers. Your profile is visible to signed-in users subject to your visibility settings. The organiser of an event you join and people you are introduced to receive relevant profile information, event answers, published contact details and introductions. Currently, leaving the public-email field blank uses your account email as your contact email; provide a different contact email if you do not want to share that address.
Assistants you connect. If you connect an AI assistant, such as Claude or ChatGPT, to your account, that assistant acts for you: it receives what you ask it to fetch or change through the connection, including event details, your answers, your introductions and, only when you ask for a particular person, the published contact details of someone you were introduced to. The assistant's provider processes that information under its own terms and privacy notice, as your choice and not as our processor. You can see what connected assistants changed and disconnect them at any time on your account page.
Recipients may retain information already shared. Changing visibility or deleting your account does not automatically erase their copies. We handle required recipient notifications and assist with relevant requests. Organisers are responsible for their own subsequent uses of participant information.
Service providers and external APIs. Hosting, storage, authentication, email, AI and other backend API providers receive information needed for the functions they perform. Their retention, logging, review and permitted-use arrangements can differ. We do not make a blanket promise that every provider operates without retaining information or using it for training or improvement. Where a provider processes personal information on our behalf, we remain responsible for putting required processing agreements and safeguards in place. Any provider acting for its own purposes must be identified with those purposes and the applicable legal basis in section 5. External login providers also process information under their own privacy notices.
Legal and business purposes. We may disclose information when legally required or lawfully necessary to investigate abuse, protect people or defend rights. A merger, acquisition or business transfer may involve relevant information, subject to confidentiality, lawful safeguards and any required notice.
We do not sell personal information.
5. International processing and providers
Information may be processed where we and our providers operate, including outside your country. Where required, we use recognised transfer safeguards, such as applicable adequacy decisions or approved contractual safeguards, with additional measures as necessary. Contact us for information about the safeguards or a copy where available.
Current recipients. These providers receive information today, each for the function named:
| Provider | Function | Information | Where | Role and terms |
|---|---|---|---|---|
| Cloudflare, Inc. | Hosting and edge network, application database and file storage, fallback AI model and the header image model (Workers AI) | Everything the service stores; for Workers AI, the AI inputs listed in section 3 | Global edge; data stored in the European Union by default | Processor under the Cloudflare Customer Data Processing Addendum; Workers AI states that inputs are not used to train its models |
| Hetzner Online GmbH (Hetzner Inference) | The AI model that assesses connections and writes introductions | The AI inputs listed in section 3 | Germany | Processor. The inference service is offered as a beta without a service commitment; we have no verified retention or training commitment beyond Hetzner’s published terms, which we identify as an unverified aspect |
| Modal Labs, Inc. | Standby AI model server (used only if the primary provider is unavailable) | The AI inputs listed in section 3, only while in use | European Union region | Processor under Modal’s terms; no retention of prompts on our server beyond the request |
| Amazon Web Services EMEA SARL (Simple Email Service) | Sending service email | Your email address and the content of the message | Frankfurt, Germany | Processor under the AWS Data Processing Addendum |
| Google LLC and LinkedIn Ireland | Sign-in, if you choose it | The identity information shown on their permission screens | Per their notices | Independent controllers for the sign-in itself, under their own privacy notices |
Potential providers. The following are providers or model services under consideration, not a statement that they currently receive your information. Before any of them receives participant information we identify the receiving entity, processing countries, role, safeguards, retention and training terms for that exact service, and update this section. Provider documents can change; listing a document does not confirm that an agreement has been executed or a safeguard enabled.
| Provider or service | Official processing documents | Scope and verification notes |
|---|---|---|
| DeepSeek | Open Platform API Terms; Privacy Policy (cdn.deepseek.com/policies) | Direct API use is distinct from models hosted elsewhere. A standalone processing addendum was not verified; the public privacy policy is not a substitute for it. |
| Mistral AI | Data Processing Addendum; Privacy Policy; zero-data-retention documentation (legal.mistral.ai, docs.mistral.ai) | Depends on the contracted plan and enabled controls. Zero data retention requires approval and has endpoint exclusions; it is separate from training preferences. |
| OpenAI | Data Processing Addendum; API data controls and retention (openai.com/policies, developers.openai.com) | API and business terms apply to backend use. Endpoint-specific retention and optional controls must be verified; consumer ChatGPT settings say nothing about the API. |
| Anthropic | Data Processing Addendum (anthropic.com/legal) | Incorporated into Anthropic’s commercial terms. When Claude is reached through another platform, that platform’s agreement and processing chain also apply. |
| OpenRouter | Data Processing Agreement; Privacy Policy; data collection and provider controls (openrouter.ai) | Both the gateway and the actual receiving model host, including fallback routes, would need to be identified and verified. |
| Qwen via Alibaba Cloud Model Studio | Alibaba Cloud Data Processing Addendum; Model Studio privacy notice (alibabacloud.com) | Qwen is a model family, not a recipient. We run Qwen models through Hetzner and previously Modal; these documents would apply only to Alibaba Cloud hosting. |
| Cerebras | Privacy Policy; Service Terms; Trust Center with a Data Processing Agreement (cerebras.ai) | The full applicable agreement was not verified. The published privacy policy includes statements about input and output retention whose scope would need checking. |
| OpenCode (opencode.ai) | Privacy Policy; Zen gateway documentation (opencode.ai) | A development tool, distinct from its hosted services and the selected model provider. Not a recipient of participant information. |
6. Retention and deletion
We keep account information while your account remains active. We retain profiles, event answers and introductions for as long as needed to provide the associated account and event features, then delete or anonymise them when that purpose ends. Email delivery logs are retained for 90 days.
We retain support and security records according to the time needed to resolve the issue, investigate abuse or establish and defend relevant claims. Legal records are kept for applicable mandatory retention periods. Retention decisions consider the purpose, sensitivity, ongoing need and legal requirements.
You may request account deletion at any time. We delete or anonymise information in our systems that is no longer needed, subject to lawful retention. Deletion from AgentsIntro does not mean that every external recipient’s copy, log or backup is erased immediately. Retention and deletion depend on the recipient’s role, applicable legal requirements and the processing arrangements described in section 5. We do not promise simultaneous deletion across all external services. We handle applicable deletion requests, communicate required instructions or notifications to processors and other recipients, and take other steps required by applicable law and our provider arrangements. This does not limit your statutory deletion rights.
Backups. We do not run scheduled backups of our own database at present. A manual export taken before a release is deleted once the release has settled, at the latest 30 days after it was taken. Our hosting provider’s own point-in-time recovery keeps database history for up to 30 days.
Beta content may be reset or removed during testing. Real personal information remains subject to this policy throughout beta.
8. Your choices and rights
Depending on applicable law and the relevant conditions, you may access or correct your information, request deletion, restrict or object to processing, request a portable copy, withdraw consent, and challenge qualifying automated decisions. Withdrawing consent does not affect earlier lawful processing. You may also complain to a competent data-protection authority; for us that is the Estonian Data Protection Inspectorate (Andmekaitse Inspektsioon).
Where we rely on consent, you may withdraw it at any time by contacting hello@agentsintro.com or using any available consent controls. We stop the processing covered by that consent, including future disclosures relying on it. If an affected feature cannot lawfully operate without that processing, it may no longer be available to you. Withdrawal does not retrospectively invalidate earlier lawful processing or remove applicable rights concerning information already shared.
Contact hello@agentsintro.com to exercise your rights. We may verify your identity and will respond within the applicable legal deadline. You can also update available profile and visibility settings within your account.
The service is intended for adults aged 18 or over. If you believe a child has provided personal information, contact us so we can investigate and take appropriate action.
9. Policy updates
We may update this policy when our practices change. We will update the date and notify you of material changes as required. Any new processing requiring consent will begin only after that consent is obtained.
Contact
Tideseed OÜ · hello@agentsintro.com · contact form