Vidya Terms of Service and Data Practices

By creating an account or using Vidya (the “Service”), you agree to these Terms. If you use the Service on behalf of an organization, you represent that you have authority to accept these Terms for that organization.

If you do not agree to these Terms, do not use the Service.

You must be legally able to enter into these Terms under the laws of your location. You may not create a new account if Vidya has previously suspended or permanently banned your account.

Vidya is an AI-powered data analysis and research assistant.

Depending on the features you use, Vidya can:

• Analyze uploaded spreadsheets and other datasets
• Plan and execute data analyses
• Generate code to analyze data
• Produce written findings and charts
• Answer questions about uploaded documents
• Conduct live web research
• Use custom assistants configured by users, referred to as “subagents”
• Connect to external services you authorize (for example, Slack or Notion) and use that service’s tools on your behalf, subject to the approval settings described in Section 5
• Draft and, with your explicit confirmation, send email on your behalf

Vidya relies on third-party artificial intelligence and search providers to provide some of these features. The information those providers may receive, and how Vidya handles your data before and after processing, is described below.

Most features require an account. You are responsible for activity conducted through your account and for keeping your account credentials secure.

You must provide accurate information when creating an account and must notify Vidya if you believe your account has been accessed without authorization.

Test accounts

Certain features, such as chat and subagents, can be tried without first creating an account. If you use one of these features while signed out, Vidya automatically creates a temporary “test account” for you in the background, with no separate sign-up step, button, or confirmation shown — and no email address, password, or other personal information required.

A test account:

• Is isolated from other users’ accounts
• Has its own chat history, memory, and usage limits, which are more restricted than a registered account’s
• Cannot be accessed by other visitors
• Is automatically and permanently deleted within 24 hours of creation

Returning from the same browser on the same calendar day reuses that day’s test account and its existing usage quota rather than creating a new one. A different browser or a new calendar day creates a new test account.

To prevent abuse, Vidya may limit how many test accounts can be created from the same network within a given period. If that limit is reached, you may try again later or create a regular account.

These Terms apply to a test account for as long as it exists, including the Acceptable Use and Safety Enforcement provisions below.

Vidya’s current data-handling practices are described below.

Chat history and conversational memory

Chat messages and conversational memory are the primary exceptions to Vidya’s temporary handling of uploaded content.

Conversational memory consists of a short, paraphrased summary of a conversation. It is not a word-for-word transcript. It is used to help Vidya retain relevant context in future conversations.

You can delete chats and subagents through the Service. Deleting a chat removes its associated conversational memory. Deleting a subagent removes the memory associated with that subagent.

No model training

Vidya does not use your data to train its own artificial intelligence models or third-party models.

Uploaded datasets and documents

Uploaded datasets and documents are processed temporarily in server memory. They are not saved to a permanent database or retained as permanent files.

Uploaded documents used for document question-answering are processed for your session and are not permanently indexed or stored.

How dataset analysis works

Vidya generally processes data analysis requests in two stages.

Stage 1: Planning and code generation

Before analysis code is generated, Vidya receives a description or preview of the dataset, such as column names and data types.

Personal information in the preview is replaced or removed where detected. This may include:

• Names
• Birthdates
• Government identification numbers
• Social Security numbers
• Email addresses
• Phone numbers
• Physical addresses
• Other identifying information

The purpose of this step is to allow Vidya to understand the structure of your data without exposing unnecessary personal information.

Stage 2: Generating the answer

After the analysis runs, Vidya generates the final answer from the resulting data. This may include personal information when it is necessary to answer your question accurately.

For example, if you ask which customers generated the most revenue, the answer may include those customers’ names.

This information may become part of your saved chat history. Social Security numbers, credit card numbers, and International Bank Account Numbers are removed from final responses. Other information, including names, email addresses, phone numbers, and addresses, may appear when it is directly relevant to the requested analysis.

Third-party AI providers

Vidya may send information needed to process a request to third-party AI providers. The provider used for a particular request may vary.

Vidya applies the same data-handling rules described above regardless of which provider processes the request. Third-party provider policies may change over time, so you should review their current terms and privacy policies if that is important to your use of the Service.

Live web research

When you use live web research, Vidya sends search terms derived from your request to third-party search providers. The search terms may not exactly match the wording of your original question.

Those providers may receive:

• The search terms submitted
• The pages visited or retrieved as part of the research
• Information necessary to return search results

Vidya does not control the retention or use of information by third-party search providers. Review their current privacy policies before submitting confidential or sensitive information through live web research.

External service connections

You may connect Vidya to an external service that supports the Model Context Protocol (MCP), such as Slack or Notion, by signing in to that service and authorizing Vidya through its own login screen (OAuth). Vidya never sees or stores your password for that service.

A connection requests only the permissions (“scopes”) the service’s own authorization screen presents, and Vidya only uses the specific tools you choose to enable for that connection.

An action that merely reads information may run automatically. By default, an action that changes something outside Vidya — for example, posting a message — is held for your explicit approval before it runs, and is never carried out silently.

You may choose to turn off that approval step for a specific tool, individually or all at once for a connection, so it runs automatically going forward. This is always a deliberate choice you make and can reverse at any time — Vidya never turns on automatic approval for you, and a tool you haven’t approved this way still requires your approval before every run.

You may disconnect a service at any time. Disconnecting immediately and permanently erases its stored access credentials. Using a connected service remains subject to that service’s own terms and privacy policy.

Email sent on your behalf

You may ask Vidya to draft an email. Vidya shows you the recipient, subject, and body, and sends nothing until you explicitly confirm it.

A confirmed email is sent through Resend, a third-party email-delivery processor, from Vidya’s own verified sending address. Your own registered email address is set as the reply-to address, so any reply goes directly to you rather than through Vidya.

A daily limit applies to how many emails can be sent through Vidya on your behalf. See Section 4.

Conversational memory

To preserve context between conversations, Vidya may create a short, paraphrased summary of a conversation. The summary is not a word-for-word transcript.

Conversational memory is associated with your account and is separated from other users’ information. It remains under your control and can be removed by deleting the relevant chat or subagent.

You may not use the Service to:

• Bypass, disable, or interfere with Vidya’s safety systems
• Use adversarial prompts, prompt injection, or repeated rewording to evade a safety restriction
• Harass, defraud, threaten, or endanger another person
• Upload information you do not have the legal right to use
• Violate another person’s privacy or rights
• Attempt to obtain restricted service information
• Conduct activity that violates applicable law
• Interfere with the operation, security, or availability of the Service
• Use a connected external service, or Vidya’s email-sending feature, to take an action you are not authorized to take, or to send unsolicited bulk messages or spam

You are responsible for ensuring that your use of the Service complies with all applicable laws, contracts, professional obligations, and organizational policies.

Every request may be subject to an automated safety review before it is processed. Automated reviews may be inaccurate. The Safety Enforcement section explains how Vidya handles repeated or serious violations.

Vidya tracks how many times an account has triggered a safety decline, to identify patterns of misuse. The declined request text and the reasoning behind each decision are logged separately, for a limited time — see Section 4.

A single ambiguous or mistaken request does not automatically result in account suspension.

Possible enforcement steps

Automated review flag

If an account exceeds an internal threshold of safety declines, it may be flagged for review. This flag does not automatically suspend or restrict the account.

Warning

If a review indicates a pattern of deliberate misuse, Vidya may issue a warning and provide a way for you to respond.

Suspension or termination

If misuse continues after a warning, or if conduct is sufficiently serious, Vidya may:

• Suspend or permanently terminate the account
• Delete associated Service data
• Cancel an active subscription
• Take other measures reasonably necessary to protect the Service or other users

Except for an automated review flag, enforcement decisions are made through human review. If you believe an enforcement action was incorrect, contact Vidya using the information in Section 12.

Paid plans are billed on a recurring basis through Vidya’s payment processor.

If your plan allows usage beyond its included quota, additional usage is billed at the rate displayed when you subscribed or otherwise authorized the overage.

You may cancel at any time. Unless otherwise stated, cancellation takes effect at the end of the current billing period, and you may continue using the paid features until then.

If your account is suspended or terminated for violating these Terms, your active subscription may be cancelled. You will not continue to be charged for a Service that you can no longer access.

Any refunds are governed by the refund terms presented at the time of purchase or by applicable law.

You retain ownership of the data, documents, and other content you submit to Vidya.

You represent that you have the rights and permissions necessary to upload and process that content.

You grant Vidya a limited, temporary license to process your content solely as necessary to:

• Provide the Service
• Analyze your data
• Generate and return responses
• Maintain chat history and conversational memory when those features are enabled
• Protect the security and integrity of the Service
• Investigate voluntarily submitted support or issue reports

This license does not transfer ownership of your content to Vidya. Vidya does not obtain a general right to use your uploaded content for unrelated purposes or to train artificial intelligence models.

Content is handled and retained according to the practices described in Sections 4 and 5.

The Service is provided on an “as is” and “as available” basis.

AI-generated analysis may be inaccurate, incomplete, misleading, or based on an incorrect interpretation of your data. You are responsible for reviewing and verifying any output before relying on it, particularly when the output may affect legal, financial, medical, employment, academic, operational, or other consequential decisions.

Vidya does not guarantee that:

• The Service will always be available
• The Service will be uninterrupted or error-free
• Results will be accurate or complete
• The Service will meet a particular purpose or requirement
• Charts, calculations, citations, or research results will be free from errors
• The Service will be compatible with every dataset, document, device, or workflow

Vidya is not a substitute for professional advice.

To the extent permitted by law, Vidya disclaims all warranties not expressly stated in these Terms.

Vidya may update these Terms as the Service changes.

When reasonably possible, Vidya will provide advance notice of material changes. The updated version will identify its effective date.

If you continue using the Service after the updated Terms take effect, you agree to the revised Terms. If you do not agree to the changes, you must stop using the Service and may cancel your account or subscription.

For questions about these Terms, the Service, or Vidya’s data-handling practices, contact: jw07@navi-os.cc