# Hemory Hemoryは常時稼働するリスニング・メモリーツールです。現実の会話を文字起こしして検索可能なアクティビティに整理し、内蔵および外部のAIエージェントで利用できます。会議や音声メモなどを要約、回答、文書に変換します。 - Findings24: - ウェブサイト: - 料金: Freemium - カテゴリー: AIミーティングアシスタント, 音声テキスト化, AIナレッジベース ## Hemoryとは？ Hemory is a listening and memory tool that captures spoken experiences, transcribes them, and organizes the results into searchable activities. It is intended to help users retain context from meetings, conversations, voice notes, and other real-world interactions. The captured material can be queried in Hemory’s built-in Q&A agent or shared with compatible external agents through the Model Context Protocol (MCP). Those agents can use the memory as source material for follow-ups, reports, journal entries, product documents, and other generated outputs. ## Hemoryでできること ### 100-language transcription Transcribes spoken sessions in 100 languages and produces timestamped, speaker-tagged, AI-polished transcripts. ### Activity segmentation and summaries Automatically separates captured audio into activities and creates type-aware summaries and minutes. ### Speaker recognition and search Recognizes speakers by voice where available, supports transcript search, and lets users replay individual transcript lines on the listening device. ### Built-in Q&A agent Answers questions about captured memories, including questions about decisions, commitments, and earlier conversations. ### MCP for external agents Connects Hemory to compatible MCP clients so external agents can search memories and use them to generate documents and other outputs. ### Manual and scheduled listening Listening can be started manually or scheduled for defined periods. Voice-activity detection excludes silence, gaps, and background noise from listening-time calculations. ## 利用シーン ### Meeting follow-ups Search previous discussions for commitments, owners, and dates, then use an agent to prepare a follow-up for a later check-in. ### Work reporting Provide a month of captured work conversations to an MCP-connected agent and generate a report deck with progress, takeaways, and next steps. ### Nightly journaling Schedule a recurring workflow that turns notable moments heard during the day into a brief journal entry. ### Product requirements drafting Combine customer interviews and internal brainstorming sessions as source material for a structured product requirements document. ### Mobile voice capture Use the iPhone, Apple Watch, or Android app to capture conversations and voice notes that need to be searchable later. Android support requires version 10 or later. ## よくある質問 ### What devices support Hemory? The site lists iPhone, Apple Watch, and Android. Android requires version 10 or later. Web, macOS, Windows, Linux, and self-hosted versions are listed as coming soon or planned. ### How does Hemory count listening time? Voice-activity detection counts periods when someone is speaking. Silence, gaps, and background noise do not consume the listening allowance. ### Can Hemory connect to other AI agents? Yes. Hemory supports MCP connections to compatible clients, with examples including Claude Code, Codex, Gemini, Cursor, VS Code, OpenClaw, and Hermes. ### Where is raw audio stored? Hemory states that raw audio is not stored in the cloud. It is kept briefly for transcription and then permanently deleted; audio on the listening device remains local and can be managed or deleted there. ### What happens after a listening allowance runs out? Listening can continue, but new audio is no longer uploaded for processing until additional allowance is available. Previously generated memories remain available. ## クイック情報 - **Category**: Productivity and AI memory - **Primary workflow**: Listen, transcribe, search, and generate from spoken context - **Transcription**: 100 languages - **Agent connection**: Model Context Protocol (MCP) - **Supported devices**: iPhone, Apple Watch, and Android - **Listening modes**: Manual and scheduled