Megabook TECNO AI Introduction

1. Product Introduction

TECNO Al is an Al platform based on an on-device large model, offering a series of A applications designed to enhance business productivity and dailylife efficiency. Since the on device model is deployed locally on the PC, most functions can be used completely offline, ensuring full data security.

2. Supported System Languages

•Chinese system: Supports mutual translation between Chinese and English.

•English system: Supports mutual translation between Chinese and English.

•Russian system: Supports mutual translation between English and Russian.

3. Launch TECNO AI

When launching the app for the first time, the built-in on-device largemodel requires a startup time of approximately 20 seconds, depending on the computer'sCPU and memory. For continuous use, it is recommended to keep the app running in the background to avoid startup delays upon reopening.

Tips: For optimal performance, connect to a power source during use.

4. Homepage

The homepage provides quick access and functional descriptions of all TECNO AI modules. Click on a function card to enter directly.

•Supports text input for on-device model Q&A, local translation, and Deepseek model Q&A.

•Supports voice input for on-device model Q&A, local translation, and Deepseek model Q&A.

Tips: Deepseek is only supported on MEGABOOK S14/S16.

5. AI Meeting (On-Device Model)

AI Meeting is an all-in-one intelligent meeting tool that supportsreal-time or offline multilingual transcription and translation. It automatically distinguishes speakers and generates meeting summaries and to-do lists. It works securelywithout an internet connection and supports direct audio recording and transcripticon within third-party apps.

5.1 Real-Time Meeting

Real-Time Meeting is a tool for real-time voice transcription and translation. It can record local and online audio across any third-party app.

•Supports real-time voice transcription and bilingual translation.

•Supports system audio transcription and bilingual translation.

•Supports meetings of up to 1 hours; recordings are automatically saved when the time limit is reached.

•Allows pausing recordings and customizing meeting names during sessions.

•After stopping recording, audio and transcript files are automatically saved locally and can be viewed in the history.

5.2 Document Transcription

Document Transcription is a tool for transcribing audio/video files into text.

•Supports audio file transcription and bilingual translation for files up to 60 minutes in AAC/FLAC/M4A/MP3/WAV AV formats.

•Supports video file transcription and bilingual translation for files up to 60 minutes in AVI/MKV/V/MOV/V/MP4/WMA/WMV formats.

•Allows custom task names during transcription.

5.3 Speaker Differentiation

Differentiates speakers in the audio after transcription.

•Supports up to 10 speakers.

•Allows custom speaker names.

5.4 AI Summary

Generates AI-powered summaries of transcribed content after transcription.

•Extracts key meeting points.

•Generates meeting to-do lists (if applicable).

•Supports editing AI summaries; modifications in the source language are automatically translated if bilingual tasks are enabled.

•Supports copying AI summary content.

5.5 History

View all historical meeting records.

•Supports one-click transfer of meeting content to the knowledge base.

•Allows searching historical records by date.

•Supports deleting/batch deleting historical records.

•Supports editing file names.

•Supports viewing saved file locations.

6 AI Gallery (On-Device Model)

AI Gallery is a tool for intelligently managing photos and videos. By placing photos/videos in the designated AI Gallery directory, users can search for them using natural language.

6.1 Video and Photo Data Import

First-time use requires importing video and photo data. Steps:

1. Click "Open Root Directory" or the "Open Storage Location" icon to access the storage path.

2. Place photos/videos to be imported into the root directory.

3. AI Album will process the data at ~10 images/second until all data is imported.

4. Search functionality can be used directly after import.

Tips: If the album data doesn’t update, switch to the homepage and reopen the album function.

6.2 AI Gallery

AI Gallery categorizes photos, transforming management from "cluttered" to "smart and efficient." AI organizes photos/videos by time, location, and faces.
Categories include:

•All: Collection of all photos and videos.

•Favorites:User-favorited photos and videos.

•Time:Organized by year/month/day.

•Characters:Organized by facial similarity; names can be edited.

•Location:Organized by geographic coordinates.

6.3 AI Image Search

Search for photos/videos using natural language descriptions or uploaded images, improving efficiency compared to traditional memory-based searches.

•Text Search:Describe the target photo/video with keywords (e.g., "panda eating bamboo"). Supports multi-dimensional searches with time and location (e.g., "photos taken in Shenzhen in 2023").

•Image Search:Upload an image to find similar ones. Supports JPG/JPEG/PNG formats (<5MB).

•Voice Search:Use the microphone in the search box; manually stop recording and edit transcribed content.

•Ella Voice Assistant:Wake Ella and say, "Find photos/videos of XXXX."

•TECNO Phone Integration:Connect TECNO phones via PC Manager, enable photo backup to AI Album, and use text/image search after transfer.

Tips: After transferring data to AI Gallery search results may be incomplete until the system finishes importing data. Wait 1-5 minutes before searching (~10 ilmages/second processing speed).

7. Local Knowledge Base

The Local Knowledge Base is a database stored on the PC, not hosted on cloud servers. It provides a personal knowledge source for the on-device large model, enabling it to answer questions based on private data.
Think of it as a private, digital expert brain tailored to your AI.

7.1 Knowledge Base Management

Manage files stored locally for the on-device large model.

7.2 Local Knowledge Q&A

The local Q&A function retrieves and answers questions based on your knowledge base data, works offline, and ensures data security.

1. Ask questions about successfully uploaded documents, focusing on document content.

￮Supports stopping searches.

￮Supports regenerating responses.

￮Lists reference sources; documents can be opened for review.

•Supports Q&A for single/multiple attachments, focusing on attachment content.

Attachment Upload Steps:

a. Click "Upload Attachment."

b. Select files from your computer (single/multiple; DOCX/DOC/PDF/TXT formats, DOCX/DOC/PDF ≤30MB, TXT ≤10MB, max 10 files).

c. Enter your question and click "Send."

7.3 Cloud Model Q&A (S14 Models Only)

The cloud-based smart model provides fast, comprehensive Q&A support but requires an internet connection.

•Supports free Q&A with second-level responses.

•Supports saving data; click "Add to Knowledge Base" to store output locally.

Tips: Supports switching between local and cloud models; history is shared between models.

7.4 Offline Translation

Offline Translation works without internet, translating text or uploaded documents between supported languages (see system details).

•Text Translation: Limited to 3000 characters.

•Document Translation: Supports DOCX/TXT formats (DOCX ≤10MB, TXT ≤3MB, max 1 file). Preview translated documents directly after completion.

Document Translation Steps:

a. Click "Upload Attachment."

b. Select the document to translate.

c. Click "Send" and wait for translation to finish.

8. Ella

Ella is an AI voice assistant that can be activated offline to perform voice tasks.

8.1 Preconditions for Use

TECNO AI must be open to wake and use Ella.
Ways to activate the voice assistant:

•Click the Ella icon in the app’s top-right corner.

•Click the "TECNO Ella" card on the homepage.

•Press the physical F12 key on the keyboard.

•Enable "Auto-start Ella when opening the app" in settings.

8.2 Hotkey Wake-Up

Wake and open Ella by pressing the F12 key on Tecno laptops.

1. If TECNO AI is closed, pressing the hotkey opens the app and Ella.

2. If TECNO AI is open but Ella is closed, pressing the hotkey opens Ella.

3. If both TECNO AI and Ella are open, pressing the hotkey wakes Ella.

8.3 Voice Wake-Up

Wake Ella by saying the wake phrase: “Ella, Ella.”

TIPS:Waking up Ella in a noisy environment may fail. It is reconmended that you use a hotkey to wake her up.

8.4 Command Execution

The current version of Ella understands and executes commands.
After waking Ella, the assistant can execute the following commands:

•Open AI Homepage/Open AI Home

•Open AI meeting

•Open real-time transcription

•Open File Transcription/Open Offline Transcription

•Open history files

•Open AI gallery

•Search for photos with xx elements from a location on a certain date

•Open Knowledge Base/Open AI Q&A / Open GPT

•Search xxxx

•Online Search xxxx（Only S14）

•Open AIPPT

•Exit

If Tecno PC Manager is installed, Ella can also execute:

•Open PC Manager /Open Manager

•Switch to power saving mode

•Switch to balance mode

•Switch to performance mode

•Open capture voice in room

•Open remove ambient voice

•Open clear voice

•Open Device Connectivity / Open Connectivity / Connect phone /Connect pad

•Open hardware information

•Open repair drivers

•Open optimize startup items

•Clean large files

•Open pop-up blocking

9. AI PPT

Simplifies PPT creation using AI technology, enabling users to quickly generate high-quality presentations.

How to Use: Click AI PPT to open in a browser (due to regional restrictions, this feature currently requires a browser).

9.1 Experience and Login

￮Supports login with Google or other email accounts; login via email and password.

9.2 From Conversation

•Enter the PPT topic in the dialog box, set parameters (e.g., pages, audience, language, style), and upload reference documents for more professional content.

9.3 From Imported Document

•Upload Word, PDF, TXT, Xmind, Freemind, or Markdown files, or paste Markdown/long text for generation.

9.4 Import from Google Drive or From Web

•Generate by pasting Google Drive links or webpage URLs.

•Generate Outline

￮Edit outline lines or click "Refresh" to regenerate after generation.

•Template Selection

￮Edit outline lines or click "Refresh" to regenerate after generation.

•Download and Export

￮After generation, download directly, share, or edit in the online editor.

•Online Editing

￮Change templates, edit outlines, insert elements, and adjust settings in the editor.

￮Click text to edit directly.

￮AI Image helps quickly create PPT images.

•History List

￮Click "History" to access and continue editing previous works.

10. Settings

10.1 Voiceprint

Enable voiceprint recognition so Ella only responds to registered voices.

•Supports adding one voiceprint.

•Supports enabling/disabling voiceprint recognition.

•Supports custom voiceprint names.

Steps to Add Voiceprint:

a. Enable voiceprint recognition and click "Add Voiceprint."

b. Follow prompts to record voiceprint.

10.2 Image Search Mode

Three search modes to adjust AI Album search precision.

•Accurate Mode: Strictly matches all search conditions; most relevant results.

•Normal Mode: Intelligently expands keywords for a balance of richness and accuracy.

•Comprehensive Mode: Broadens range to include results matching any condition.

10.3 Ella

Enable "Auto-start Ella when opening the app" to launch Ella simultaneously with the app.

10.4 Hotwords

"Hotwords" are keywords or phrases prioritized for recognition, improving accuracy for important terms.
Steps to Add:

1. Open Hotwords and click "Add Hotword."

2. Enter keyword/phrase and save.

10.5 Storage Location

Displays the storage path for AI Album photos/videos; click to jump to the location.

11. Q&A

11.1 Why does the program sometimes respond slowly?

￮ Because the antivirus software is performing real-time scanning or intercepting and analyzing the app's behaviors, files, or network activities in thebackground.

￮ Solution: Add your App's main program and data directory as"exclusions" in all antivirus software.

11.2 Why do we need a knowledge base? What problems does it solve?

Large language models (e.g., ChatGPT) have broad general knowledge but three limitations:

￮Unaware of private/unpublished information (e.g., company financial reports, internal documents).

￮Knowledge cutoff: Training data isn’t updated beyond a certain point.

￮May "hallucinate" or generate plausible but incorrect answers. The local knowledge base addresses these by letting the AI "read" and use your exclusive data, making it more professional, accurate, and personalized.

11.3 Why does the hotkey for Ella not respond or respond slowly?

￮The hotkey is the standalone F12 key; check if Fn is enabled anddisable it before pressing F12.

￮If the app is closed, pressing the hotkey starts the app first, therElla.

11.4 Why are some hotwords not recognized correctly after adding?

￮ Hotwords aren't 100% guaranteed; real-time transcription prforitizes semantics over hotwords.

11.5 Why does Ella' s wake-up success rate decrease after adding aa voiceprint?

￮Voiceprint recognition depends on recording quality; record prompt text clearly in a quiet environment.

11.6 How to enable microphone permissions if real-time transcription prompts for access?

￮Go to System Settings > Privacy & Security > Microphone, and enable access for TECNO AI.

11.7 Why are there no search results after placing photos/videos in the designated location?

￮After first-time placement, the system extracts features; wait ~1 minute before searching.

￮Some photo/video formats may be unsupported (images: JPG/JPEG/PNG/BMP/WEBP; videos: MP4).

￮Search keywords may be incorrect; refine descriptions if needed.