🚀 Platform Updates | July 18, 2025

[ashlynn-brown]

🚀 Platform Updates | July 18, 2025

This week we rolled out major enhancements across AI tooling, backend performance, and developer experience, including new memory capabilities, deep research tooling, and seamless Google Docs integrations.

🤖 AI Mentor Enhancements

Google Docs Integration

• Mentors can connect to Google Docs for smarter, document-based chat interactions, bringing even more flexibility to knowledge ingestion and context-aware responses.

Memory Toolkit Introduced

• We've added a robust memory framework to give AI mentors persistent session context across chats.
• This lays the foundation for more adaptive, conversational tutoring and support.

Deep Research Mode Released

• A new "Deep Research" tool and LLM mode has been added to empower agents with multi-step, action-driven exploration capabilities.
• Automatically enables advanced reasoning chains and auto-selects the right model when activated.

Mentor Chat Pipeline Refactored

• A new HttpLLMRunnerConsumer endpoint allows direct HTTP-based chat interactions, reusing the refactored LangFlow consumer across both HTTP and WebSocket layers.

Expanded Credential Management

• New delete endpoints were added for both LLM and integration credentials, streamlining management and improving security hygiene.

Credit Enforcement Improvements

• Credit access now consistently validates has_credits across endpoints, blocking anonymous usage on mentors and ensuring fair resource distribution.

Screen Sharing + Call Controls Enhanced

• LiveKit availability is now checked before issuing call credentials.
• The screen sharing system has been modularized into a selectable tool for more flexible configuration.

💻 Backend & Catalog Enhancements

Archived Course Handling

• Catalog APIs now support filtering enrollments by archived courses, improving querying flexibility and historical data analysis.

Resource Query Optimization

• Resource retrieval endpoints now support query parameters, enabling more refined data requests from external systems.

System Cleanup and Consolidation

• Deprecated APIs and legacy platform metadata endpoints were removed to reduce overhead and streamline the platform's surface area.

Automatic Credit Allocation

• New users now automatically receive platform credits upon signup, ensuring immediate access to AI mentor functionality.

🧰 DevOps & Platform Infrastructure

Unified Credit Enforcement for Custom Domains

• Logic updates ensure that custom domains and admin users are no longer inadvertently restricted by credit validation.

Core Platform Enhancements

• Updates to support platform-wide SDK improvements, notification service upgrades, and cross-service credential syncing were deployed.