Recall AI

In today's information age, capturing and retaining knowledge across various sources presents a significant challenge. Recall addresses this through AI-powered content summarization and structured knowledge management. This technical overview explores how the platform captures diverse content, manages user information securely, and supports long-term knowledge retention.

Company Overview and Technology

Recall is an AI platform that helps users capture and retain knowledge from diverse content sources, including YouTube videos, articles, and podcasts. The company's technology automatically summarizes this content and organizes it into a structured knowledge base, while using spaced repetition and active recall techniques to aid memory retention.

The platform's core functionality revolves around content summarization and organization. When users encounter interesting content online, they can add it to their personal knowledge base through the browser extension or in-app search features. Supported content types include YouTube videos, articles, PDFs, and blog posts, allowing users to capture knowledge from a wide range of sources.

Upon adding content, Recall generates a structured knowledge card that users can further edit and format using the platform's built-in text editor. Each knowledge card contains four main components: a link to the original content, a set of tags for organization, a bullet-point summary generated by Recall's AI, and a series of linked cards that highlight connections between related information.

To support long-term retention, Recall employs two key strategies. The spaced repetition system reviews content at increasing intervals over time, drawing from psychological research on effective learning. Additionally, the platform incorporates active recall techniques by allowing users to test their knowledge through generated questions based on the saved summaries.

The company prioritizes user privacy and data security throughout its technical infrastructure. Data is primarily stored locally on user devices for maximum privacy, while the platform performs necessary queries and database operations directly on these devices. For added security, data is backed up and synced through a Google Cloud data center located in Belgium's europe-west1 region. Users retain full control over their data, with options to export it as markdown files or permanently delete their accounts.

The platform offers flexible payment options, including bank/credit cards, Apple Pay, and various digital payment systems. Subscribers benefit from access to advanced features, while the free tier allows 10 AI-generated summaries per month. All users maintain the ability to manually add and edit content, even after canceling their subscription.

Using Recall for Content Capture

Recall provides users with two primary methods for adding knowledge cards to their personal knowledge base: the browser extension and in-app search.

The browser extension, available for Chrome and Firefox, requires installation and pinning to the browser toolbar for optimal use. To create a knowledge card, users click the Recall extension icon in the browser toolbar, followed by the save option to add the content to their knowledge base. Supported content types include YouTube videos, PDFs, blog posts, Wikipedia pages, news articles, and recipes.

The in-app search feature allows users to create knowledge cards through both URL pasting and search term entry. Supported sources include Wikipedia, Wikidata, and Google Knowledge Graph. The search process generates knowledge cards for selected search terms, with additional functionality for creating cards on specific topics such as people, places, movies, or organizations.

Each knowledge card in Recall consists of four main components: a link to the original content, tags for organization, a summary generated by the AI, and linked cards that establish connections between related information. The card's summary is designed to capture key points of the content and can be broken into sections with appropriate headings. Users have full control over the summary through a built-in text editor that supports formatting options.

The platform categorizes linked cards by data type, including organizations, people, and products. Each keyword extraction generates linked cards, with previews displayed when hovering over blue links. The linked card system helps users discover relationships in saved content and resurface past information when new content mentions the same topics.

Content Management and Organization

The knowledge cards in Recall are structured with four main components: a link to the original content, tags for organization, a summary generated by the AI, and linked cards that establish connections between related information. Each card can be fully edited using the platform's built-in text editor, which supports formatting options for better organization.

Users organize cards using a hierarchical tagging system where tags are assigned based on content type and source. This system enables nested tagging, allowing for detailed categorization of information. The platform suggests existing tags and allows users to create new ones, with multiple tags stored per card to facilitate diverse categorization.

The platform's knowledge graph organization helps users discover connections in their saved content and resurface relevant information. When new content mentions keywords already present in the knowledge base, related cards are automatically connected. These connections are visualized through blue links within the content, with previews shown when hovered over. The linked cards are categorized by type (organizations, people, products) and may display additional cards in the knowledge base that mention the same topic.

Technical Infrastructure and Data Security

The technical infrastructure supporting Recall's operations prioritizes both performance and security. Data storage follows a principle of least privilege, with queries and database operations executed directly on user devices whenever possible. This approach minimizes data transmission and enhances local processing efficiency, though it means that certain operations must be handled through the secure Google Cloud data center located in Belgium's europe-west1 region.

For users seeking to export their data, the platform provides comprehensive options through its Settings menu. Via the Export feature, users can save their knowledge base as markdown files, maintaining the original structure and content. This capability enables seamless transition to alternative applications while preserving the organization and relationships established within Recall's knowledge graph.

In the unlikely event of platform failure, users retain their ability to protect and manage their data through several means. The project's open-source nature allows users to host Recall independently if the official service becomes unavailable. Additionally, the platform documentation outlines a clear process for account deletion, which removes data from both local devices and server infrastructure, ensuring complete data removal upon request.

The platform's payment infrastructure supports diverse financial systems, including bank/credit cards, Apple Pay, Google Pay, Cash App, WeChat Pay, and Link. Users enjoy flexibility in managing their subscriptions through the main app menu, where they can cancel at any time while maintaining access to their existing content and knowledge graph connections. While AI-generated summaries become inaccessible after cancellation, users retain full capabilities for manual content addition and organization.

Subscription and Usage Features

The platform offers a free tier allowing users to create 10 AI-generated summaries monthly, with unlimited manual content addition and editing capabilities. Paid subscriptions provide access to advanced features while maintaining these fundamental usage options.

Recall supports multiple payment methods, including bank/credit cards, Apple Pay, Google Pay, Cash App, WeChat Pay, and Link. Users can manage their subscriptions through the app menu, canceling at any time while retaining access to their existing knowledge base and saved content.

Following cancellation, users lose access to AI-generated summaries but maintain full capabilities for manual content addition and organization within their personal knowledge base. The platform's technical architecture supports continued content management and organization even after subscription termination.

For users seeking to preserve their content beyond the subscription period, Recall provides several data management options. Through the Settings menu, users can export their entire knowledge base as markdown files, maintaining the original structure and relationships established within the platform's dynamic knowledge graph.