logologo
  • User Guide
  • English
    • English
    • 简体中文
  • User Guide
    Install
    FAQ
    Next pageInstall

    #FlowyAIPC User Guide

    Dictionary

    FlowyAIPC is a large-model AI tool that runs locally on your computer, enabling AI to seamlessly integrate into your daily work through safer on-device inference. From meeting minutes, note organization, smart screenshots, and knowledge base management to comprehensive capabilities across writing, learning, operations, and productivity powered by 760+ AI agents—everything is processed, controlled, and stabilized on your device. No cloud dependency, no data leakage. Empowering everyone with an instantly accessible, private, and secure intelligent work system.

    #Applicable Configuration

    Info

    FlowyAIPC: Currently supports AI chips (excluding ARM architecture chips), 16GB or more RAM, and Windows 11 operating system.

    Official download links: flowyaipc.cn flowyaipc.com

    Installation and Activation

    We currently offer two login methods:

    1.Email Login

    login-1

    2.Scan the QR code with WeChat

    login-2

    #Feature Overview

    #Privacy Mode

    Once enabled, all data within the software—including meeting minutes, notes, and content from AI-powered features—is stored locally without uploading to the cloud.

    #1.First-Time Setup

    When first enabling Privacy Mode, a download prompt will appear requiring installation of four auxiliary privacy tools:

    - ollama supports running local models

    - bge-m3 vector model supports knowledge base Q&A

    - minicpm-v multimodal model supports knowledge base image retrieval (small model)

    - funasr model is a local ASR model supporting meeting notes and recording features in Privacy Mode

    private mode-1

    private mode-2

    #2.Multiple Openings

    After successfully downloading the four required models, simply reactivate Privacy Mode to begin using them immediately.

    Once Privacy Mode is activated, all related files and interaction records within the software—including Q&A content—will be strictly stored on your local device, ensuring the confidentiality and security of your data.

    private mode-3

    #Home

    The home screen of this app integrates recent records, to-do list management, and real-time weather. It supports uploading photos and files within interactive dialog boxes to ask questions and receive precise answers.

    home

    #Dialogue

    Users can ask questions to the AI agent to assist with various tasks.

    dialogue

    #Meeting Minutes

    To use the meeting minutes feature for the first time, click the “Create Meeting Minutes” button to initiate and begin recording the meeting content.

    meeting-1

    After using the meeting notes feature multiple times, the interface will display a meeting history log, enabling users to review and reflect on past sessions at any time. The history log includes real-time recordings, real-time transcription text, complete meeting minutes, and mind maps. Certain content within this log supports exporting, facilitating users in preserving and sharing materials.

    meeting-2

    During the meeting minutes process, the system will use an ASR (Automatic Speech Recognition) model to transcribe speech into text in real time. Simultaneously, users can take meeting notes on the right side to record key information and edit meeting topics and attendee details at any time. After the meeting concludes, the system will automatically organize and generate standardized meeting minutes by integrating the complete audio recording.

    meeting-3

    Upon conclusion of the meeting, the system will instantly display the meeting summary, key action items, and a mind map of meeting content on the right side. All generated materials can be promptly and securely saved locally. Additionally, users may choose to regenerate the meeting summary as needed to achieve optimal results.

    meeting-4

    meeting-5

    #Notes

    FlowyAIPC Notes supports manual text input and uploading of Markdown files (such as travel plans, work schedules, etc.), while providing comprehensive basic text editing tools (including headings, font styles, underlining, italics, hyperlinks, and image insertion).

    notes-1

    The note-taking feature combines the dual characteristics of a comprehensive digital notebook and a lightweight sticky note app: it supports inserting important information such as photos, screenshots, and files, while also enabling users to quickly jot down ideas and set reminders anytime. This fully addresses diverse needs ranging from in-depth documentation to rapid memo-taking. Users can also export files, with current support for docx, PDF, and Markdown formats.

    notes-2

    Additionally, through the “AI Q&A” feature on the right, users can directly pose intelligent inquiries about the current document content.

    notes-3

    #Knowledge Base

    The Knowledge Base enables users to upload business materials, which are automatically parsed by the system. Users can then directly pose questions to the AI, and the system will provide precise and professional responses based on the knowledge base content. When business content changes, only the relevant files need to be updated.

    knowledge base-1

    The knowledge base analyzes your documents, processes, FAQs, and other materials to develop business understanding capabilities. This enables it to reduce repetitive work, ensure consistent and professional responses, and prevent information bias.

    Currently supports uploading files in formats such as docx, markdown, and txt.

    knowledge base-2

    #Model Store

    The Model Store offers local large model options. These models are characterized by their significant storage space requirements on the device, longer download times, and the need for users to maintain a stable network connection during the download process to ensure download integrity and success rates.

    The Model Store features two types of models:

    - Local Models (run on GPU)

    - WML Models (run on NPU)

    model shop-1

    model shop-2

    The model store will adapt to your computer's hardware configuration.

    - A blue button indicates the model is available for download and deployment;

    - A gray button indicates the model cannot be downloaded due to your computer's configuration.

    After clicking download, please wait for the large model to complete downloading. Once the model download is successful, you can select and click the downloaded model in the “Switch Model” section. After loading is complete, you can begin using it.

    model shop-3

    model shop-4

    #Agents

    Upon entering the Intelligent Agents interface, users can filter by category tags or utilize the search bar to precisely locate target agents. This platform has aggregated over 750 intelligent agents for users to select and utilize. Additionally, the platform supports users in customizing and creating their own exclusive agents.

    agent

    #Model Switching

    Select the “Model Switching” option to view three primary model types:

    - Local Models: Models downloaded and deployed by users via the Model Store, excluding WML models.

    - Cloud-based Large Models: These models require no download and can be enabled for immediate use.

    - WML Models: Using these models requires devices running Windows 11 version 26200 or later. These models support execution on the NPU (Neural Processing Unit).

    model switching

    #Settings

    The Settings page allows users to configure the following:

    1. Language Switch: The software currently supports both Simplified Chinese and English interfaces.

    2. Font Size Adjustment: Modify the display font size in the chat interface.

    3. Automatic Update Channel: Offers three update modes for users to choose from:

    - Disable Updates: The software will not update. (Given the software's rapid iteration pace, we do not recommend enabling this channel to avoid missing important feature updates.)

    - Stable Channel: Software updates will occur one week after test versions are released. Features in this channel are relatively stable and reliable.

    - Development Channel: Enabling this channel provides access to more features but may cause instability during software use. Please choose with caution.

    settings-1

    The page contains information about the software, including its version number and a description of the software.

    settings-2

    #User Feedback

    Users can promptly obtain support and submit feedback through the following channels:

    1. Email Contact: The page provides two dedicated email addresses for users to contact us promptly regarding issues.
    2. Online Feedback Submission: Users can directly ask questions or submit suggestions about the software via the “Submit User Feedback” feature.
    3. Customer Support: Hover your mouse over the designated area and scan the displayed QR code to add customer support for inquiries.

    feedback

    #Floating Ball

    The Floating Ball menu includes Quick Notes (notes), Meeting Captions (meeting minutes), OCR Smart Screenshots (smart screenshots), and To-Do Lists (to-dos generated from meeting minutes, plus the to-do list on the right side of the homepage). Open it to start working instantly. The floating ball offers convenient one-tap operation to begin tasks immediately.

    boll

    #Smart Screenshot:

    Copy: Extract text from the screenshot and copy it to the clipboard

    Ask Flowy: Extract text from the screenshot and perform AI-powered Q&A

    Translate: Extract text from the screenshot and translate it, supporting seven languages

    Text Extraction: Extract text from the screenshot

    Insert Notes: Extract text from the screenshot and insert it into notes (Flowy's native note-taking feature, which can be saved to your knowledge base)

    #Meeting Captions:

    This software supports two meeting modes: main window and floating caption window. Additionally, when watching foreign-language videos, users can enable the caption style feature for real-time translation into Chinese and other languages. The system currently supports seven different languages for translation.