Technical Guide: Organizing AI Conversations with Memoriq

Practical Summary and Purpose

Memoriq is an encrypted, private memory vault designed to store and organize conversations from major AI providers including ChatGPT, Claude, Gemini, and Grok. It addresses the limitations of native provider chat histories, which are often fragmented, difficult to search, and lack robust organization tools. By utilizing browser-side encryption, Memoriq ensures that the server operator never gains access to your plaintext data, such as message bodies, titles, or source URLs. This tool is intended for users who treat AI interactions as personal knowledge work—such as debugging sessions, legal research, or project planning—and require a secure, portable, and searchable repository for their long-term records.

Prerequisites and Local Development

To self-host Memoriq, you must have a development environment capable of running a Laravel application. The documented requirements include PHP 8.3 or higher, Composer for dependency management, Node.js with npm for frontend asset compilation, and a MySQL database. The application architecture relies on Laravel Sanctum and Fortify for authentication, Vue and Pinia for the frontend, and Vite for asset bundling. If you choose to host the application yourself, you must ensure your server environment is configured to handle these dependencies and that you have established a secure database connection as defined in your environment configuration.

git clone https://github.com/memoriqme/memoriq.git
cd memoriq
composer install
npm install
cp .env.example .env
php artisan key:generate

Documented Workflow and Extension Integration

The primary method for capturing conversations is through the official Memoriq browser extension, available via the Chrome Web Store. The workflow begins with the user logging into the Memoriq web application, after which the extension connects using a scoped Sanctum token. Once the user unlocks their encrypted vault within the extension, they can trigger a capture on a supported AI chat page. The extension performs best-effort DOM extraction to pull the conversation content, encrypts the header and body locally within the browser, and uploads only the resulting ciphertext to the server. Users can save chats via a floating button on the page or through the extension popup, where they can assign the conversation to a specific project.

Privacy Model and Data Security

Memoriq is built on a strict privacy-first model where the server acts only as a storage layer for encrypted payloads. The server stores encrypted conversation headers, bodies, key envelopes, and salts, alongside operational metadata like row IDs and timestamps. Crucially, the server does not store any plaintext information, including project names, source URLs, or message content. Encryption is handled entirely in the browser using a locally unwrapped master encryption key. Because of this design, if a user loses both their encryption password and their recovery key, the data becomes permanently inaccessible, as Memoriq cannot perform password resets or data recovery for encrypted vaults.

Limitations and Fallback Mechanisms

As a beta project, Memoriq has specific limitations regarding capture reliability and media support. Because AI provider interfaces frequently update their DOM structures, the extension may occasionally fail to extract content correctly. In such instances, users are encouraged to use manual fallback methods, such as pasting an AI reply directly into the application. Furthermore, the current version does not preserve rich media files like images, audio, or video, focusing instead on reliable text-based conversation storage. While a PWA share target exists for mobile devices, it is currently under development and may only capture links rather than full conversation content. Users should verify their saved data to ensure critical information was captured successfully.

Choosing When to Use Memoriq

Memoriq is best suited for users who prioritize data sovereignty and long-term organization of their AI-generated insights. It is an ideal solution for professionals who need to maintain a searchable, private archive of debugging logs, writing drafts, or complex research that spans multiple AI platforms. By centralizing these conversations into a single, portable vault, users can avoid the pitfalls of provider-specific history limitations. Whether you choose to use the hosted service or self-host the application, the project's commitment to verifiable privacy and open-source transparency makes it a viable tool for those who want to maintain control over their digital knowledge base without relying on third-party SaaS databases for plaintext storage.