# Privacy Policy

> What Trilocore stores, where it is stored, how long backups and logs are kept, and which subprocessors are in the path.

[← Back to Trilocore](https://trilocore.com/)

Last updated: August 2026

This policy explains what information Trilocore ("we", "us") holds, where it is stored, how it is protected, how long backups and logs are kept, and which third parties are involved. It covers this website (trilocore.com) and the Trilocore Platform — the Contract IDE, the Architecture Explorer, and the Auditing IDE, together with the fork environments, analysis engine, and the workspace, API, reporting, and collaboration services that support them, and any additional features we make available under the Trilocore Platform in future.

This document is under legal review. It is written to describe our engineering practice accurately, but it has not been reviewed by counsel.

## 1. What we collect

### Account identity

The email address associated with your account, and — if you sign in with Google or GitHub — the account identifier that provider returns to confirm who you are.

### API keys

Programmatic access uses API keys. We store only a SHA-256 hash of the key. The key itself is shown once, at creation, and is never written to our systems, so it cannot be recovered later.

### Workspace content

Whatever you put into a workspace: contract source code, development notes, and the files that make up a project. Source code is held in a content-addressed object store. It is never stored as bytes in the database — the database holds only a pointer to the object.

Our workspace and engagement service stores no source and no bytecode at all. It holds reference metadata only: chain, contract address, and function selectors.

### Session and execution artifacts

The state produced when you run an analysis is stored so that a result stays reproducible: on EVM networks that is forked chain state, session transactions, and snapshots; on Solana it is the captured account set, instruction data, and the resulting account diffs.

### Findings and attestations

The findings you record against a contract, and the signed attestations issued when work is sealed for handoff.

### Usage metering

Token counts, for quota and billing. Metering never records the text of a prompt or of a response.

### Service logs

Our API gateway logs no request or response bodies. Its audit record carries identifiers for the call and a salted pseudonymous user id, not content.

## 2. What we do not do

- We never ask for private keys, seed phrases, or signing access — and you should not give them to any tool.
- We do not train models on customer content. Our training corpus is public on-chain bytecode, published bounty findings, and synthetic data.
- We do not sell personal data.
- We run no analytics, no advertising, and no third-party tracking — on this website or in the product.

## 3. Where it is stored, and how it is protected

Production runs in a single AWS region. Workspace source is held in a content-addressed object store on encrypted network storage, never as rows in the database. Analysis and provenance artifacts are held in a separate object store under its own dedicated AWS KMS key. Database volumes, block storage and the shared file system are each encrypted at rest with AWS KMS. The artifact store's policy rejects any request that is not made over TLS.

This website is a static site served by GitHub Pages. We do not operate its web server and we load no third-party scripts on it. The only scripts are a few lines written inline in the pages themselves; there is no analytics, tag manager, or tracking of any kind.

## 4. Retention

Workspace content — contract source, development notes, analysis artifacts and findings — is kept for as long as the workspace exists. There is no automatic expiry on it; it is removed when you remove it, or on request. The windows below are the separate, automatic ones that apply to our own operational copies.

| What | Retained for |
| --- | --- |
| Database point-in-time recovery | 30 days — recovery to any moment inside the window |
| Database backup objects | 45 days — deleted from storage no later than this |
| Application logs | 30 days |
| Cloud audit and storage access logs | 90 days |
| Shared file-system backups | 30 days |

## 5. Subprocessors

These are the third parties that can come into contact with your data. Some are always in the path; others only if you choose a feature that uses them.

| Provider | Purpose | When |
| --- | --- | --- |
| Amazon Web Services | Hosting, object storage, key management, secrets, backups, and the database | Always |
| Amazon SES | Transactional email | Always |
| GitHub Pages | Serves this website, and therefore receives the requests your browser makes to it | Every visit to trilocore.com |
| Google Fonts | Serves the web fonts. Your IP address reaches Google as part of that request | Every page load |
| Google | OAuth sign-in | Only if you sign in with Google |
| GitHub | OAuth sign-in; publishing a workspace to a repository you control | Only if you sign in with GitHub, or opt into publishing |

We may also disclose information where the law requires it.

## 6. AI-assisted features

Trilocore's AI-assisted features are not currently active. No customer content is sent to a model provider. If we enable them, the processor involved will be named in the subprocessor table above before the feature is switched on.

## 7. Cookies

This website sets no cookies. The product application sets first-party cookies to keep you signed in. There are no analytics, advertising, or cross-site tracking cookies anywhere in Trilocore.

## 8. Security

We apply technical and organizational measures appropriate to the sensitivity of the material involved, and we treat unreleased contract source as confidential by default. No system is completely secure; we work to safeguard your data and to respond promptly to any incident. Our [Trust Center](https://trilocore.com/security) describes the model in more detail, and our [security.txt](https://trilocore.com/.well-known/security.txt) carries the disclosure contact.

## 9. Changes

We may update this policy. Material changes will be reflected in the "Last updated" date above, and a new subprocessor will appear in the table above before it enters the path.

## 10. Your choices

Write to [hello@trilocore.com](mailto:hello@trilocore.com) and we will tell you what is held for your account, correct anything inaccurate, and provide a copy of your workspace content in a portable form. If you want content removed, say so and we will handle it with you directly; note that operational backups above roll off on their own schedule rather than immediately.

Formal data-processing terms are not published yet. If you need a data processing agreement, a data residency commitment, or an incident notification commitment before you can adopt Trilocore, contact us and we will work through it rather than point you at a page.

## 11. Contact

Questions about this policy, or about the information held for your account, go to [hello@trilocore.com](mailto:hello@trilocore.com). Security reports go to [security@trilocore.com](mailto:security@trilocore.com).

---

[Canonical HTML version](https://trilocore.com/privacy)