Privacy Policy

1. Introduction

FIDGARD, Inc. ("FIDGARD," "we," "us," or "our") is committed to protecting the privacy of our users and the confidentiality of client trust accounting data. This Privacy Policy explains what information we collect, how we use it, who we share it with, and your rights regarding your data.

This policy applies to all users of the FIDGARD platform, including law firm administrators, attorneys, and staff members. Given the sensitive nature of legal trust accounting data, we hold ourselves to the highest standards of data protection.

2. Information We Collect

2.1 Account Information

When you create an account, we collect:

• Name and email address
• Organization name
• Password (stored as a BCrypt hash; we never store plaintext passwords)
• Role within your organization (Admin, Attorney, Staff)
• Timezone preference

2.2 Trust Accounting Data

In the course of using the Service, you may enter or upload:

• Trust account details (account names, bank information)
• Client matter names and personal ledger records
• Transaction and payment records, including deposits and checks
• Payee and vendor information
• PDF documents (checks, closing statements, payment instructions)
• Reconciliation data

This data belongs to you. We process it solely to provide the Service and do not use it for any other purpose.

2.3 Bank Data (via Plaid)

If you connect a bank account through Plaid, we receive:

• Account balances (current and available)
• Transaction history (posted and pending transactions)
• Account identifiers (masked account numbers)

We store Plaid access tokens in encrypted form. We do not store your bank login credentials. Plaid's handling of your data is governed by Plaid's End User Privacy Policy.

2.4 Accounting Data (via QuickBooks Online)

If you connect QuickBooks Online, we access:

• Chart of accounts (to map trust accounts)
• Accounting records created in FIDGARD and posted to QuickBooks
• Linked-account balances and previously posted records used to verify exports and prevent duplicates
• Company information (name, realm ID)

QuickBooks OAuth tokens are stored encrypted. We post accounting records created in FIDGARD when you initiate a sync or as configured in your settings. Supporting reads enable account mapping and posting verification.

2.4A Practice-Management Data (via Clio Manage)

If your firm connects Clio Manage, FIDGARD receives information permitted by the integration and the connecting Clio user's access. Depending on the workflow, this includes contact names, email addresses and phone numbers; matter names, identifiers and descriptions; invoice numbers, dates, status, totals and balances; account names and identifiers; and matter account balances. This may include confidential client information.

We store imported records in your firm's FIDGARD workspace for client and matter linking, invoice review, trust-balance comparisons and related accounting review. Cached source responses may include associated fields returned by Clio. We also store connection identifiers, authorization credentials and synchronization status. Newly stored Clio access and refresh tokens are encrypted at the application level.

The current Clio app requests read-only Accounting, Billing, Contacts, Matters and Users access, without write or webhook permission. Imports save data in FIDGARD; they do not send a bill, post a payment or change Clio records. Reads can occur when you request an import or when configured background maintenance refreshes cached records. If imported information is included in AI workflows or shared through an optional ChatGPT or Claude connection, the disclosures in sections 2.5 and 2.5A apply.

You can disconnect Clio in FIDGARD. This clears locally stored OAuth credentials and connection-specific caches and links. It does not undo transactions, erase required accounting or audit history, or delete every historical record containing a Clio identifier. Disconnecting in FIDGARD does not guarantee that the authorization is revoked in Clio. For retained-data requests, contact privacy@fidgard.com; the retention limits in this policy apply.

2.5 Documents and Records Processed by AI

The following two AI workflows send data to OpenAI's API, and we want you to know exactly what leaves our infrastructure.

Document extraction. When you upload a document — or one arrives at your firm's intake inbox — we first read any embedded text layer on our own infrastructure. No third party receives the file at that step. We then send rendered images of the document pages to OpenAI, together with the extracted text, so the model can read amounts, dates, payees and account details in their original layout. Scanned and image-only pages are transcribed the same way. This means the page image itself — the bank statement, the closing statement, the check — is transmitted to OpenAI.

AI assistance over your trust records. Ask the Books, Trust Review explanations and Smart Inbox send trust accounting records to OpenAI. Ask the Books runs your question against your books and sends the results — including ledger entry descriptions and memos, client and matter names, payee names, dates and amounts. Reconciliation matching is not in this list: it is rule-based and deterministic by design, so it reaches no third party.

Applying to both:

• Not used for training; held briefly by OpenAI for abuse monitoring. Every request sets OpenAI's store=false parameter, which keeps the prompt and response out of the stored history OpenAI would otherwise make retrievable, and OpenAI's API data usage policy states that API inputs are not used to train their models. That is not the same as OpenAI holding nothing: OpenAI ordinarily retains API inputs and outputs for up to 30 days for abuse and misuse monitoring, then deletes them. Thirty days is the usual limit rather than a guaranteed one. OpenAI may hold data beyond it where the law requires, where retention is needed to protect OpenAI's services or third parties, or where its systems flag content — images and uploaded files included — for manual human review. Treat 30 days as the normal case and longer as possible in those situations.
• Redaction is narrow, and we will not overstate it. One path — the structured extraction step for uploaded documents — applies best-effort masking of Social Security numbers, dates of birth, driver's licence numbers and passport numbers before the text is sent. This does not guarantee that all sensitive data is removed. The workflow does not provide general redaction of bank account or routing details, which may be needed for extraction and transmitted. Page images are never redacted — anything visible on the page reaches OpenAI as it appears, including signatures and account details. Smart Inbox and Ask the Books apply no redaction at all; their text reaches OpenAI as stored. Assume that anything in a document you upload, or in a ledger memo, can be sent verbatim.
• Where your documents actually live. Uploaded documents are stored in our own Amazon S3 buckets, encrypted at rest, for as long as your firm keeps them — Rule 1.15 record retention depends on that. AWS is our infrastructure provider for it, as listed in the subprocessor table below. OpenAI receives a copy only for the processing described above; we do not ask it to store one.

The configured production OCR path uses OpenAI. If the OpenAI OCR module cannot be imported, the PDF text extractor can fall back to AWS Textract when available, sending document images to AWS for OCR. An unavailable OpenAI service or a failed request does not by itself switch this extractor to AWS.

If your firm's obligations do not permit client documents or matter data to be sent to OpenAI, contact us before uploading documents or enabling AI features.

2.5A Optional ChatGPT and Claude Connections

Where the FIDGARD connection is available in your ChatGPT or Claude account, you may choose to connect it. You sign in to FIDGARD and select the firms the connection may read. FIDGARD checks your current role, firm access and connection grant on read requests. The public connection can find selected firms, list their trust accounts, return account setup and balance information, and look up recorded checks and historical withdrawals with their clearing evidence. It cannot create or change accounting records, import a workbook, connect a bank or initiate a payment.

Depending on the request and account setup state, information returned to the service you choose can include firm, trust-account, client and matter names and identifiers; bank names; book and client-ledger balances; opening and cutover status, amounts and related dates; deposit and outstanding-check details, descriptions, memos and references; bank-statement filenames and document identifiers; setup completion information; and bank connection status, balances and timestamps already held by FIDGARD. This may include confidential client information. Bank information is labeled cached, stale or unavailable. The account setup read makes no live bank request, and a FIDGARD book balance is not necessarily a current bank balance.

Check lookup may also return payee and client names, check and reference numbers, dates, amounts allocated to client ledgers, recorded check and clearing status, clearing dates or other stored evidence, and identifiers linking allocations of a combined check. Historical withdrawals whose payment method is unknown are labeled separately. Check-clearing results reflect records held by FIDGARD; the lookup does not contact a bank and is not a live bank confirmation.

The receiving service handles information under its own account settings and terms. FIDGARD does not control whether information already returned remains in a conversation or how that service retains or uses it. The store=false setting used for FIDGARD’s separate OpenAI API processing does not establish your ChatGPT account’s data-handling settings and does not apply to Claude.

Without optional renewable access, a FIDGARD connection lasts no more than ten minutes and also ends if the originating browser session expires or is revoked. Where these renewable options are offered, you explicitly choose during authorization between read-only access for seven days or ongoing read-only access until you disconnect it. The seven-day option ends seven days after authorization; use and background credential renewal do not extend that deadline. The ongoing option has no scheduled expiry, but manual revocation, security changes or loss of required firm access can stop it as described below. Neither renewable option disconnects because it is unused, and both can continue after passive browser sign-in expiry. Access tokens still last no more than ten minutes; the connected service can renew them only while the connection remains valid. Renewal does not add firms or permissions. Existing connections keep their original access mode and deadline; they are not automatically upgraded to either renewable option.

When you open FIDGARD during the final 24 hours of a valid seven-day connection, the dashboard and Settings can show an expiry reminder. Ongoing connections have no scheduled-expiry reminder. Keep connected opens instructions; it does not extend access. Complete a new authorization in the connected service and review the selected firms and access duration to start a new connection. The reminder is shown in FIDGARD, not promised as a notification from ChatGPT or Claude. Without new authorization, a seven-day connection ends at its original deadline.

Use FIDGARD Settings → Connected apps to review or revoke a connection. Revocation or expiry stops subsequent reads through that grant. Password changes, explicit revocation of the originating browser session, account security invalidation, loss of required firm access or archival of a selected firm can also stop access earlier. Passive browser-session expiry alone does not end an explicitly renewable connection. Disconnecting in ChatGPT or Claude is a separate action and may not immediately revoke a still-active FIDGARD grant. Expiry or revocation does not delete information already returned to the other service or change accounting records.

2.6 Usage and Analytics Data

We automatically collect:

• IP addresses (for security and audit logging)
• Login timestamps and session duration
• Feature usage patterns (which pages and features you use)
• Browser type and device information

2.7 Payment Information

Subscription payments are processed by Stripe. We do not store credit card numbers or bank account details for payments. Stripe collects and processes payment information under Stripe's Privacy Policy. We receive only a Stripe customer ID and subscription status.

3. How We Use Your Information

We use your information to:

• Provide the Service: Manage trust accounts, process transactions, generate reports, and run reconciliations
• Maintain security: Authenticate users, prevent fraud, and maintain audit trails
• Process payments: Bill your subscription and manage your account
• Communicate with you: Send transactional emails (password resets, billing notices, system alerts)
• Improve the Service: Analyze usage patterns to fix bugs and improve features
• Comply with law: Respond to legal processes, enforce our Terms, and protect rights

We do not:

• Sell your data to third parties
• Use your trust accounting data for advertising
• Share your data with other FIDGARD customers
• Use your documents to train AI models
• Mine your data for insights unrelated to providing the Service

4. Data Sharing and Subprocessors

We share data with the service providers listed below as necessary to provide FIDGARD. If you separately choose to connect ChatGPT or Claude, the information returned to that service is described in section 2.5A. That user-directed connection is distinct from FIDGARD’s OpenAI API processing listed in the table.

Clio Manage is an optional integration you choose to connect. FIDGARD exchanges authorization information and read requests with Clio to retrieve the practice-management data described in section 2.4A.

We may also share information if required by law, court order, or governmental authority, or to protect the rights, property, or safety of FIDGARD, our users, or the public.

5. Data Security

We implement multiple layers of security to protect your data:

• Encryption in transit: All data transmitted between your browser and our servers uses TLS encryption
• Encryption at rest: Database and file storage are encrypted at rest
• Token encryption: Third-party API tokens (Plaid, QuickBooks) are encrypted before storage using application-level encryption
• Password security: Passwords are hashed using BCrypt with salting
• Access controls: Organization-level data isolation ensures no cross-tenant data access; role-based permissions restrict actions within organizations
• Audit logging: All data modifications are logged with user, timestamp, IP address, and before/after snapshots
• PII redaction: System logs automatically redact sensitive data patterns (SSNs, credit card numbers, API keys)
• Immutable records: Cleared trust accounting transactions cannot be edited or deleted, preserving the audit trail

6. Data Retention

Trust accounting records are subject to professional responsibility rules that typically require retention for 5 or more years after the end of representation. Our retention practices reflect this:

• Active accounts: All data is retained for the duration of your subscription
• Canceled accounts: Your Data is available for export for 90 days after cancellation. After 90 days, data may be deleted unless retention is required by law
• Audit logs: Retained for the life of the account and not subject to deletion
• System logs: Retained for 90 days for operational purposes (PII redacted)
• Payment records: Retained as required by tax and financial regulations

We do not automatically purge trust accounting records. If you need records deleted, contact us and we will work with you to balance your request against applicable retention obligations.

7. Your Rights

Depending on your jurisdiction, you may have the following rights regarding your personal data:

• Access: Request a copy of the personal data we hold about you
• Correction: Request correction of inaccurate personal data
• Deletion: Request deletion of your personal data, subject to legal retention requirements
• Portability: Request your data in a structured, machine-readable format
• Restriction: Request that we limit processing of your data in certain circumstances
• Objection: Object to processing of your data for specific purposes

To exercise any of these rights, contact us at privacy@fidgard.com. We will respond within 30 days.

Note: Certain trust accounting data may not be deletable if retention is required by professional responsibility rules or applicable law. We will explain any such limitations in our response to your request.

8. California Privacy Rights (CCPA)

If you are a California resident, you have additional rights under the California Consumer Privacy Act (CCPA):

• The right to know what personal information we collect, use, and disclose
• The right to request deletion of your personal information
• The right to opt out of the sale of personal information — we do not sell personal information
• The right to non-discrimination for exercising your privacy rights

9. Children's Privacy

The Service is intended for use by legal professionals and is not directed at individuals under 18 years of age. We do not knowingly collect personal information from children. If you believe we have collected data from a minor, please contact us immediately.

10. Cookies and Tracking

We use the following cookies and similar technologies:

• Authentication cookies: HttpOnly cookies maintain signed-in sessions, including a separate partner session where applicable.
• Security cookie: csrf_token, a random value your browser sends back on state-changing requests so we can reject forged ones. It is readable by JavaScript on our own pages, because the application must echo it as a request header, and it has a maximum age of seven days that may be refreshed during use. It supports the security of requests to the Service.

These cookies support authentication and request security. The application uses no advertising, analytics or attribution cookies, and loads no third-party advertising or analytics tags. It does not provide an advertising or analytics cookie-choice control. Browser storage is also used for product functions, including your display preference and cached account information.

We do not sell personal information, and we do not track you across other websites.

11. International Data Transfers

The Service is hosted in the United States. If you access the Service from outside the United States, your data will be transferred to and processed in the United States. By using the Service, you consent to this transfer.

For users in the European Economic Area (EEA), we rely on Standard Contractual Clauses (SCCs) or other approved transfer mechanisms where required by GDPR.

12. 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 at least 30 days before the changes take effect. Corrections that make an existing practice described more accurately — without changing what we do with your data — take effect when published, since delaying them would only leave a less accurate description in place. The "Effective Date" at the top of this page indicates when the policy was last revised.

13. Contact Us

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

FIDGARD, Inc.
Privacy inquiries: privacy@fidgard.com
General support: staff@fidgard.com