MemoryCrux

MemoryCrux is the unified versioned memory layer that synchronises and organises AI conversation history across multiple platforms, providing a single searchable archive of everything a user has discussed with AI assistants. In an era where professionals routinely use multiple AI tools throughout their workday, MemoryCrux ensures that valuable context and knowledge generated in those conversations is never lost or siloed.

The synchronisation layer connects to OpenAI and Claude conversation histories, importing and indexing conversations with semantic deduplication to eliminate redundant content. If a user discussed the same topic across multiple AI platforms, MemoryCrux identifies the overlapping content and consolidates it under a unified representation while preserving links to the original conversations on each platform. Staleness detection identifies conversations that reference outdated information, flagging them for review.

Versioned snapshots capture the state of the user's conversation memory at regular intervals, exposed through the Model Context Protocol for consumption by AI assistants and other tools. This means that an AI assistant can query MemoryCrux to understand what the user has previously discussed, what conclusions were reached, and what context is relevant to the current conversation, all without the user having to manually re-explain their background.

The Developer Tier provides BM25 keyword search with receipted results, allowing users to search their conversation history using natural language queries and receive results with CueCrux's standard provenance guarantees. Every search result links back to the original conversation with timestamp, platform, and context information. Higher tiers add semantic search capabilities that understand conceptual similarity, enabling queries like finding all conversations about a particular topic regardless of the specific terminology used.

Temporal reconstruction, available in premium tiers, allows users to reconstruct the evolution of their understanding of a topic over time. By tracing conversations chronologically and identifying how the user's questions and the AI's answers evolved, temporal reconstruction creates a narrative of knowledge development that can be valuable for research documentation, audit trails, and personal knowledge management.

MemoryCrux integrates with the broader CueCrux platform by feeding conversation insights into the Engine's context model. When a user asks a question on the Ask Surface, Engine can draw on MemoryCrux to understand the user's background knowledge, previous queries on related topics, and preferred level of detail. This creates a more personalised and efficient search experience where the platform learns from every interaction.

For professionals who rely heavily on AI tools, MemoryCrux transforms scattered conversation histories from a fragmented liability into a structured, searchable, receipted asset.