Hemory captures spoken conversations on mobile devices and exposes a searchable record to AI coding and assistant tools via the Model Context Protocol.
What the product does and how it is positioned
Hemory is an ambient voice memory application designed to record real-world conversations and make them accessible to AI agents.
The software processes audio locally or via streaming, organizing captured context into a searchable timeline accessible through Model Context Protocol integrations.
Source-supported ways to use the product
AI agents can query past meeting records to check agreed-upon deadlines, assignments, and specific commitments.
Connected agents can synthesize recent discussions and interviews to draft status reports, presentation decks, or product requirement documents.
AI assistants can extract highlights from recorded daytime conversations to produce recurring evening journal summaries.
The documented workflow, where available
Initiate recording manually when needed or configure a schedule to listen automatically during designated recurring hours.
Add the Hemory MCP server to a supported client to expose the memory search tool to the connected AI assistant.
Hemory functions as an ambient memory layer bridging physical spoken conversations with external AI coding assistants and autonomous agents. Rather than keeping transcripts isolated inside a standalone interface, the platform exposes a Model Context Protocol tool named search_memory to external MCP clients. This allows development environments and conversational assistants to retrieve grounding context directly from real-world meetings and discussions.
Data privacy and stream management are structured around localized audio retention. Streaming audio sent for cloud processing is discarded once transcription finishes, while physical audio files remain strictly isolated on the capture hardware. This architecture is intended to reduce multi-device cloud synchronization while still allowing users to query memory records and generate derivative artifacts through connected agents.
Checks to run with your own material and workflow
What was checked and when
Answers based on the source-checked product record
Hemory is available on iOS, Apple Watch, and Android. Versions for macOS, Windows, Linux, Web, and self-hosted environments are listed as coming soon.
Hemory works with standard Model Context Protocol clients, including Claude Code, Codex, Gemini, Cursor, VS Code, OpenClaw, and Hermes.
Raw audio is stored exclusively on the recording device and is never synced across devices. Audio processed in the cloud is treated as an ephemeral stream and destroyed immediately after processing.
Users can activate Manual mode to record only on demand, or Schedule mode to automatically listen during predefined time frames, such as standard working hours.
Hemory uses voice-activity detection to identify active speech, so that background noise, conversational pauses, and silence are separated from spoken dialogue.
