Opaline Privacy

This Privacy notice explains how Opaline Labs, Inc. collects, uses and discloses personal information in connection with the hosted Opaline service.

1. Who we are

Opaline Labs, Inc. is a San Francisco-based company that provides analytics for coding-agent sessions.

Opaline Labs, Inc.
2261 Market Street
San Francisco, CA 94114
United States

Email: evren@opaline.so

2. Scope

This notice applies to Opaline's website, hosted application, API, command-line integration, support and related hosted features.

It does not apply to a self-hosted deployment operated entirely by a Customer. If a self-hosted deployment sends information to Opaline's hosted services, this notice applies to the information Opaline receives.

The Service is intended for business use and is not directed to children.

3. Our role

Customers decide which coding-agent sessions are uploaded, who can access a workspace and how session analytics are used. For personal information contained in Customer-submitted session content and the analytics derived from it, Opaline acts as a service provider or processor for the Customer. The Customer's instructions and Opaline's obligations are described in the Data Processing Agreement.

Opaline acts as the business or controller for account administration, authentication, Service operation and security, support, communications, billing and product analytics.

4. Information we collect

Account and organization information

We collect names, email addresses, profile images, account and organization identifiers, workspace memberships and roles, invitations, authentication provider, login records, preferences, subscription information and protected API or CLI credentials.

Coding-agent session content

When a Customer submits a session, it can contain prompts, model responses, system messages, source code, files and file paths, terminal commands and output, tool calls, repository and version-control information, URLs, timestamps, model and token information, errors, and personal or confidential information appearing in those materials.

Opaline's server removes known secret formats (API keys and access tokens for common services, private keys) from every uploaded session before storing it, and the plugin's tools redact passwords, credentials in connection strings, bearer tokens and other credential-like values from any session text they return.

When the CLI captures a session inside a Git repository, it also collects repository context by default: Git state such as branch, commit, remote and file-status information, recent commit metadata, size-limited diffs of uncommitted changes, untracked text files, and agent instruction and configuration files such as AGENTS.md, CLAUDE.md, skills and MCP configuration. The CLI filters known secret formats from this content before upload, and we use it to link agent sessions to the code changes that result from them.

Workspace analytics

We create Customer-directed analytics from submitted sessions, such as session and interaction counts, timing, token and model usage, repository and project information, skill and agent activity, errors, classifications, reports and User-, project- or workspace-level comparisons.

GitHub information

When a Customer connects the Opaline GitHub App, we collect pull request and CI information for the connected repositories: pull request and commit metadata such as titles, states, the GitHub accounts of authors and timestamps, and CI runs and jobs with their workflows and results. We use it to show the outcomes of agent work, such as merged pull requests and CI failures.

Device, usage and log information

We collect IP address, browser and operating-system information, application and CLI version, requested pages and features, timestamps, request identifiers, authentication events, upload status, error codes, approximate location derived from IP address and information used to prevent abuse or investigate incidents.

Communications

When someone contacts us, joins a sales conversation or uses support, we collect contact details, the communication and attachments, and relevant account or technical information.

5. Sources of information

We receive information directly from Users and Customers; automatically from browsers, the CLI, integrations and use of the Service; from GitHub, when a Customer connects the Opaline GitHub App; from a Customer's workspace administrators; from identity providers such as Google or GitHub when selected; and from service providers that help us operate Opaline.

6. How we use information

We use personal information to:

• create accounts, authenticate Users and manage organizations, invitations and permissions;
• receive, store, display and analyze coding-agent sessions at the Customer's direction;
• provide dashboards, reports, workspace analytics, exports and public shares requested by a User;
• operate, maintain, troubleshoot and secure the Service;
• respond to support, sales and other communications;
• send authentication, operational and administrative messages;
• manage subscriptions, transactions and business records;
• understand feature adoption and improve reliability and usability;
• prevent fraud, abuse and unauthorized access; and
• comply with law and establish, exercise or defend legal claims.

We do not use substantive coding-agent session content to train general-purpose AI models, create advertising profiles or build cross-customer benchmarks. We do not use workspace analytics to evaluate a Customer's personnel for Opaline's own purposes.

7. AI processing

When a Customer uses an AI-assisted analysis feature, Opaline may send the content needed for that request to OpenAI or Anthropic. The relevant provider processes that content to return the requested analysis. Opaline does not authorize those providers to use Customer content to train their general-purpose models. To search past analyses by meaning, Opaline sends the question title and topics of a saved analysis to OpenAI to create search embeddings, only when the analysis is saved. To let a workspace search its sessions by meaning, Opaline sends the text of each uploaded session (prompts, the agent's messages, its tool calls and their output, and file contents the agent read), after Opaline's secret filter has removed recognised credentials, to OpenAI to create search embeddings when the session is uploaded. When an analysis asks yes/no questions about agent runs (run questions), Opaline sends the text of the selected runs, after its secret filter, to Subquadratic (SubQ), which returns the probability of each answer; Subquadratic does not use this content to train its models.

AI-generated analysis can be incomplete or inaccurate. Customers decide whether and how to use it and should not treat it as the sole basis for employment or other legally significant decisions about an individual.

8. How we disclose information

We may disclose personal information:

• to infrastructure, database, analytics, AI, email, authentication and support providers that help deliver Opaline;
• within a Customer's workspace according to its settings and permissions;
• to the public when an authorized User intentionally creates a public share;
• to professional advisers and auditors under appropriate confidentiality obligations;
• to comply with law, legal process or valid government requests, or to protect rights, safety and the Service; and
• in connection with a merger, financing, acquisition, reorganization or sale of assets, subject to customary protections.

Our current providers and their functions are listed on the Subprocessors page.

9. Cookies and analytics

Opaline uses cookies, local storage and similar technologies to maintain sessions, authenticate Users, remember settings, prevent abuse and measure product use.

We use PostHog for website and product analytics. On our public website, PostHog operates in cookieless mode without creating persistent visitor profiles. In the authenticated product, analytics may include User and organization identifiers, pages and features used, onboarding and authentication actions, upload results, application version, device category, broad content-size categories, normalized error codes, installation identifiers and hashed project identifiers. We configure analytics not to collect raw prompts, model responses, source code, file contents or terminal output.

Where required by law, we request consent or provide an opt-out for non-essential browser storage. Browser settings can also block or delete cookies, although doing so may affect the Service. Opaline does not currently respond to browser “Do Not Track” signals. We honor legally recognized opt-out preference signals, such as Global Privacy Control, where required.

10. Retention

We keep personal information only as long as reasonably necessary for the purposes described here, including providing the Service, maintaining business and transaction records, resolving disputes, enforcing agreements and complying with law.

• Account and organization information is retained while the account or workspace exists.
• Customer session content, workspace analytics and analyses are retained until the Customer deletes the relevant session, account or workspace, the agreement ends, or another valid instruction requires deletion. The analysis log's own timelines are in Section 18.
• When an account or workspace is deleted, its account and organization records are removed at once, and its sessions, analyses and workspace analytics are deleted within 7 days: a deletion job starts immediately, a second pass runs about 15 minutes later to catch data written while the first ran, and failed passes are retried.
• Public shares remain available until deleted, replaced or automatically expired.
• Operational logs are kept by our hosting provider for at most 30 days and are not exported.
• Product analytics events are kept as long as needed to understand and improve the product, then deleted or anonymized.
• Support emails are kept as long as needed to resolve the request and keep a record of it, then deleted.
• Residual copies in database backups are overwritten in our providers' regular backup cycles.

11. Privacy choices and rights

Depending on where a person lives and subject to legal exceptions, they may have the right to request access to, correction of, deletion of or a portable copy of personal information; learn about its collection and disclosure; opt out of sale, sharing or targeted advertising; limit certain uses of sensitive personal information; withdraw consent; object to or restrict processing; and appeal a denied request.

Opaline does not sell personal information or share it for cross-context behavioral advertising, and does not use sensitive personal information to infer characteristics about individuals.

To exercise a right, email evren@opaline.so. We may verify a request using account or contact information. An authorized agent may submit a request where permitted by law, but we may require proof of authority and verification of the individual. We will not discriminate against anyone for exercising a privacy right.

If personal information is part of Customer-controlled session content, the individual should normally contact that Customer first. We will assist the Customer as required by law.

12. California disclosures

During the preceding 12 months, Opaline collected the following categories of personal information, depending on how the Service was used:

The sources and business purposes for these categories are described in Sections 5 and 6. We disclosed these categories for the business purposes described above. We did not sell them or share them for cross-context behavioral advertising, and we do not knowingly sell or share personal information of anyone under 16.

13. International transfers

Opaline is based in the United States. Personal information may be processed in the United States and other countries where our providers operate. Those countries may have different privacy laws from a person's home jurisdiction.

For people in the European Economic Area, United Kingdom or Switzerland, our legal bases for controller processing are performance of a contract, compliance with legal obligations, legitimate interests in operating and improving a secure business service, and consent where required. Where required for an international transfer, we use an adequacy decision, certification, contractual clauses or another lawful mechanism.

Individuals in those jurisdictions may have rights to access, correct, erase, restrict, object, receive a portable copy, withdraw consent and complain to their local supervisory authority.

14. Security

We use administrative, technical and organizational safeguards designed to protect personal information. No method of transmission or storage is completely secure, so we cannot guarantee absolute security. Users should protect their credentials and promptly notify us of suspected unauthorized access.

15. Children

Opaline is a business service not directed to anyone under 18. We do not knowingly collect personal information directly from children. If you believe a child has provided personal information to us, contact us so we can review and delete it where appropriate.

16. Changes

We may update this notice as the Service, providers or legal requirements change. We will post the current version here and update its effective date. If a change materially affects privacy rights, we will provide additional notice where required.

17. Contact

Questions, complaints and privacy requests can be sent to evren@opaline.so or:

Opaline Labs, Inc.
2261 Market Street
San Francisco, CA 94114
United States

18. Opaline in ChatGPT and Codex

When a User connects the Opaline plugin in ChatGPT or Codex, it can read dashboards and analytics, session data, team information, skills, repositories and connected GitHub pull request and CI information in workspaces the User belongs to. The plugin never changes or deletes sessions. It writes only analyses (see below) and the User's own default-workspace setting for the plugin. Only the session owner and workspace administrators can open a session's contents. Every member of a workspace can see its dashboard, in Opaline and in the plugin, including usage and estimated cost per member; a member can see what it shows about them on the dashboard and Team pages in Opaline. Each member chooses which of their repositories upload sessions, through the Opaline CLI's per-repository upload settings. Workspace admins can remove members. A member can delete their account, which removes their sessions from every view and dashboard and deletes every analysis that read them; to have single sessions deleted, email evren@opaline.so.

The plugin keeps an analysis log. While a User asks a question through the plugin, Opaline keeps the plugin's searches and reads of session data, with their results, as a private draft that nobody can open, the User included; the searches and reads do not include the User's question or other conversation text. A draft that is not saved is deleted 7 days after it was started. When the plugin saves the analysis, the draft becomes an analysis in the workspace with a short title of the question, its findings and final answer, so the User and their team can find and build on it later; only saved analyses are kept. A saved analysis is not changed afterwards: a later, updated answer is saved as a new analysis that refers to the earlier one. A User can share an analysis with their workspace. Sharing shows the analysis, its findings and the stored results behind it (which can include short excerpts from the sessions that were read) to members of the workspace; workspaces that restrict results show those results only to admins and the session's owner. To make saved analyses searchable, their title and topics are sent to OpenAI to create a search index. Stored search and read results are deleted after 90 days, or after one year in shared or pinned analyses. Analyses are kept until the account or workspace is deleted. To have analyses deleted sooner, email evren@opaline.so.

When a User asks a question in ChatGPT or Codex, the results Opaline returns are sent to OpenAI, the provider of the assistant the User chose, to answer the question. OpenAI's own terms and privacy policy govern that data. This is separate from Opaline's AI-assisted analysis described in Section 7: OpenAI is not acting as Opaline's subprocessor for the plugin connection.

For the connection, Opaline stores the OAuth client registration, access tokens that expire after 1 hour, refresh tokens that expire after 7 days unless renewed, and the User's “Default workspace” plugin setting. Deleting the Opaline account deletes these connection records and settings.

To disconnect, remove the Opaline plugin in ChatGPT or Codex. To revoke all connections immediately, email evren@opaline.so.