Your first meeting, step by step.

Use it beyond meetings.

The same controls work for lectures, podcasts, interviews and personal notes. The desktop app keeps its current meeting labels.

• Start meeting: capture a live lecture, discussion or spoken note using your selected audio sources.
• Import audio: choose an audio or video file for local transcription instead of live capture.
• AI chat: enable both Meeting context: On and Transcript context: On in the sidebar, then ask for study notes, key ideas or explanations. Default summary formats focus on meetings; AI chat and custom prompts let you request another format.

Video imports transcribe the audio track. Add screenshots or other images through Meeting context config when visual information matters.

Before you start

• A Windows PC: check the requirements listed with the release you download.
• An internet connection: needed for account activation, speech model downloads and cloud AI requests.
• A microphone and a test audio source: a headset is useful for keeping speaker playback out of your microphone.
• A folder with enough free space: meeting recordings and downloaded speech models can take substantial space.
• Online AI credentials or a local AI model: online chat and summaries need your own provider credentials; local AI needs a supported GPU and an installed model. Speech recognition and speech generation use their own model selections.

1 Download and open MeetSimpler

1. Use Windows downloads on the homepage for the installer, or choose the portable ZIP from GitHub Releases.
2. Read the release requirements before downloading. The installer and portable ZIP are alternative ways to run the Windows app.
3. If you downloaded an installer, follow its installation instructions. If you downloaded a ZIP archive, extract the whole archive into a folder before running the app. Keep the included files together.
4. Open MeetSimpler.exe. The first screen asks you to sign in.

App will not open? Check security software blocks.

You are ready for the next step when the MeetSimpler sign-in window appears.

2 Sign in and activate desktop access

1. Read the Terms and Privacy Policy, then select the acceptance checkbox in the desktop sign-in window.
2. For email sign-in, enter your email and password and select Sign in. If you need an account, select Create account. Follow any email confirmation instructions, then sign in.
3. Alternatively, select Continue with Google or Continue with GitHub in the desktop app. Complete sign-in in the browser tab opened by the app.
4. When the browser reports that the desktop handoff succeeded, return to MeetSimpler. The app completes activation and continues to setup.

You can also register on the account page. Signing in on the website alone does not activate the desktop app: return to the app and complete its sign-in flow.

You are ready when the app opens the first-run meeting folder screen, or the dashboard if you have already chosen a folder.

3 Choose where your meetings are saved

1. In the welcome screen, choose Use suggested folder for the displayed location, Select folder for an existing folder.
2. Check the displayed path. Choose a location you can access easily and that has room for recordings and model downloads.
3. Select Continue. MeetSimpler opens the dashboard.

This is your meeting parent folder. Each new meeting gets its own timestamped folder beneath it. Recordings, transcripts, summaries and other meeting files are stored locally. Use Meeting folder from the dashboard to find the current meeting's files.

To change the location later, open Settings, choose the meeting parent folder and select Apply folder change. Read any relocation instructions shown by the app before continuing.

You are ready when the dashboard shows your selected meeting folder.

4 Install and select a speech model

Speech recognition, also called STT (speech to text), turns captured audio into a local transcript. Selecting an AI chat model does not install a speech model.

1. Open Settings from Meeting Control and find Speech recognition (STT).
2. Use Use CPU only for a first setup without a working CUDA environment. Use Use GPU with CUDA only when the app's CUDA checks confirm support.
3. Select Install and manage STT models.... Locate the model required for your chosen device and select Install in its Action column. The app currently uses Whisper base as the CPU default and Qwen3-ASR-0.6B as the CUDA default.
4. Keep the internet connection available while the model and its packages download. Wait for its status to show Installed. If it reports an incomplete installation, use the offered reinstall action.
5. Return to Settings. In your device's STT model selector, choose an installed model. Leave the default STT pipeline settings for your first practice session.
6. Allow the settings change to save before closing Settings. Back on the dashboard, check that the STT model is selected rather than showing None (action required).

Repair (no model download) repairs the selected model's software packages. It does not download missing model files. If the manager reports Reinstall required, use Reinstall to download and validate the model again.

Explore supported local models and how overlap decoding works.

The speech model loads when you select Start meeting. The STT control also lets you load or restart it. A first load can take longer than later uses; check the model status before expecting transcript text.

STT installation or loading failed? Check security software blocks.

You are ready when your selected speech model is installed and named on the dashboard.

5 Connect AI chat and summaries

Open Settings → AI provider config. Choose Add online model for an API connection, or Add local model after installing a local model. Saved profiles keep their own model and request limits.

1. For an online model, enter a configuration name, the provider API base URL and your API key. Use Select AI model, or Edit manually for an exact model identifier. Test connection, then Save.
2. Back in the configuration list, check the profile you want to use. Select a row and choose Edit, or double-click it, to change that profile. Saving a profile and selecting it are separate steps.
3. Review Settings → AI context settings before using real meeting content. Each task controls its transcript, instructions and supporting files. Files attached directly in chat are sent with that chat request.

Online profiles send enabled content to your chosen provider and need your own API credentials; charges may apply. Local AI runs through installed Ollama models. Account activation and renewal remain separate from either AI choice.

Set up local AI and translation

The built-in general AI catalogue includes Qwen and Gemma models for chat, summaries and script generation. HY-MT1.5 and TranslateGemma are specialist translation choices, selected separately.

In Settings → AI features, open Install and manage local AI models. Install or start Ollama, then install a model from the list. Downloads need internet access; installed models run locally. General local AI requires a supported NVIDIA or AMD GPU; specialist local translation can use CPU.

Return to AI provider config → Add local model. Name the profile, choose an installed model, Test connection and Save, then check that profile in the list. Show detailed config reveals its input, output, context, timeout and file limits.

In Settings → Local translation models, choose the active AI profile or an installed specialist translator. Local requests do not fall back to a cloud provider. For local chat, use text and a model that supports your images; attach PDFs to an online model that supports them.

Choose a local translator

Use a specialist translator for live transcript translation or selected-text and time-range translation. Chat, summaries and script generation continue to use the active AI profile.

1. In Settings → AI features, open Install and manage local AI models. Install or start Ollama and install the translation model you want.
2. In Settings → Local translation models, choose the installed specialist model. This selection is independent of AI provider config; it does not replace your chat or summary model.
3. Open Detailed analysis, choose the target language and enable Translation or request analysis of selected text or a time range. Review the result against the source transcript.

Specialist translation runs locally on CPU or supported GPU hardware and needs no online provider API key. Model downloads and account activation still need internet access. Supported language pairs depend on the chosen model.

6 Choose what the meeting will hear

1. Connect your headset or microphone and open the meeting app or test playback source you want to capture.
2. In Meeting Control, select the button beginning with Audio sources: to open Transcript source.
3. Select a microphone to capture your voice, if needed. Select either system output audio or an available app output to capture the other side of an online call.
4. Double-click an audio source row or use its checkbox to toggle it. Check Selected sources, then select Confirm.

Available application outputs can include your meeting app, browser or media player. Open the application and start playback so it appears in the source list. Select its output to capture audio from that application, or select system output to capture all audio played by the computer.

You can select up to one microphone and one output. Choosing another source in the same group replaces the previous choice.

The Voice input control is for chat voice input. Configure meeting capture through Audio sources:. Use headphones if playback is also reaching your microphone and causing repeated speech.

You are ready when the Audio sources button reflects the microphone and output you intend to use.

7 Run a short practice meeting

1. Select Start meeting. If a setup reminder appears, complete the missing settings before continuing.
2. Check that the control changes to Click to stop meeting, the recorded time advances and the STT status becomes ready. Confirm recording is on if you want saved audio.
3. Say a few clear sentences into your selected microphone. For example: “This is a practice meeting. We will review the beta on Friday. I will collect feedback before the next meeting.”
4. If you selected an output source, play a short spoken clip or have a consenting test participant speak in the call.
5. Watch Live Transcript. Temporary recognition can appear first; confirmed text should follow as the model processes the audio. Check that both your voice and the selected output appear when both are selected.
6. If text is missing, check the source selection, device mute state and STT status before recording a full meeting.
7. Select AI chat. In its sidebar, check that Meeting context: On is enabled. Select Transcript context: Off to change it to Transcript context: On if needed. This includes the meeting transcript in your chat request to the configured AI provider.
8. Type a question such as “What decisions have we made so far?” and select Send. Review the reply against the transcript.

Live chunk summaries appear after enough transcript chunks finish and AI requests complete. For a short practice session, request a final summary after ending the meeting if no live summary is due yet.

The practice worked when you can see the expected speech in the transcript and, if AI is configured, receive a chat reply.

8 Finish, review and keep the results

1. Select Click to stop meeting. Wait for the app to finish saving the meeting before requesting a final summary or opening another meeting.
2. Open AI chat and select Final summary. Wait for the request to finish. The resulting summary is saved in the meeting folder as final_summary.md.
3. Read the transcript and summary. Check names, numbers, decisions, owners and due dates against what was actually said. Correct your notes before sharing them.
4. For a specific part of the discussion, select Partial summary, choose the time range and select Generate summary.
5. Select Meeting folder to inspect your records. The confirmed transcript is saved as transcript.md; generated summaries and saved audio remain with the meeting.
6. If audio was saved, select Export full audio to produce a complete WAV file. Choose its destination and wait for export to finish.

Ending a meeting saves the available records; requesting Final summary is a separate AI action. You can retain a transcript even if an AI request fails and retry the summary once the provider connection is working.

You have completed your first meeting when you can open its folder, find the transcript and review any generated summary or exported audio.

Manage your account

Use Tab to move between controls and Enter or Space to activate the focused button. A successful email login, logout or account deletion moves keyboard focus to the newly shown heading. A background session refresh keeps your current focus. After a retryable request error, focus returns to the control you used if you have not moved elsewhere.

Use the account page to create an account or log in with email, Google or GitHub. Website sign-in is separate from desktop activation; start browser pairing from the Windows app when activating it.

Forgot your password? On the account page, enter your email and use the Forgot password? button. Open the email link, enter a new password and select Save new password. After success, log in again. If recovery fails or times out, request a new link from the account page.

For an email account, Change password asks for your current password and a matching new password entered twice. Google-only or GitHub-only accounts do not show this password form; manage those credentials with the sign-in provider.

Log out of this browser ends the current browser session. Log out on all devices ends account refresh sessions across devices. Existing desktop offline permits can remain valid until they expire; logging out does not remove local meeting files.

To permanently delete your account, choose Delete my account, review the warning, type DELETE exactly and select Permanently delete my account. Cancel closes the confirmation without deleting anything. Account deletion cannot be undone; recordings and meeting files on your computer are kept.

If an account request times out and the controls remain disabled, reload the account page. For a desktop browser sign-in timeout, return to the Windows app and start sign-in again.

Import audio or video

Import a recording instead of capturing live audio. Importing creates a new meeting folder and leaves the original media unchanged. Speech recognition runs locally with your installed speech model; optional AI analysis uses your configured provider.

1. End the current meeting and select an installed speech model. Choose Import audio in Meeting Control, then Audio file... to select an audio or video file. Video imports transcribe the audio track; they do not analyze video frames.
2. For several recordings, choose Folder of audio segments... and select their folder. Number ordinary filenames in playback order before importing. Saved MeetSimpler meeting folders use their recorded audio tracks.
3. Follow the separate progress stages for importing the file, loading the speech model and transcribing. A first model load can take longer. Select Cancel if you want to stop the import.
4. Wait for the successful import message. Use Open folder to inspect the audio and transcript, or Close to return to the dashboard. A progress bar alone does not confirm that the transcript has been saved.
5. Review the transcript, then use AI chat, Final summary or Detailed analysis as needed. The imported meeting remains paused; use Resume meeting only when you want to add new live audio.

AI chat

Open AI chat and review Meeting context and Transcript context before sending a question. Enable both to ask about the meeting transcript. Voice input can send the current draft automatically; review its settings before dictating.

Press Enter to send a message and Shift+Enter to add a new line. Clear input removes unsent text but keeps draft attachments. Remove unwanted draft files from the file list before sending.

Right-click the message box and choose Attach files, or drag supported images, PDFs or text files into the chat window. Wait for attachment processing to finish before sending. The file list distinguishes draft attachments, submitted files and files returned by the provider.

Select New chat to start another conversation in the same meeting, and confirm when prompted. Earlier saved chat sessions remain in the meeting folder; starting a new chat does not start a new meeting.

Each question and reply share a foldable block. Reopening a conversation expands its newest question first. Use Fold all or Unfold all in the header, or open individual blocks. The active block has a highlighted title.

Cancel stops that question's pending reply. Update regenerates its answer using the submitted attachments; partial and final summaries keep their original operation and time range. Remove deletes the question and reply from the saved conversation and cancels its request. Other requests continue, and local attachment files are kept.

Select a sent or received file link to open it, or Open folder to locate it. In Settings → Other settings, Render Markdown in AI Chat controls formatted answer display. Turning it off changes the display without changing the saved answer.

AI request settings and limits

Edit the selected profile in AI provider config. Online profiles use Detailed config; local profiles use Show detailed config. Save the editor to apply changes to that profile.

Online profiles estimate the text budget without counting images or PDFs, which retain provider limits. Local profiles also enforce a context ceiling and reserve space for output and images. Requests that exceed the configured limits stop without trimming the selected input.

To resolve an input-budget error, choose a shorter transcript time range or disable unneeded text context. Increase the budget only if your selected model can accept the request. Provider limits still apply to both text and media.

Request timeout defaults to 180 seconds and includes provider queueing, processing and returned files. If a request times out, check the provider or local runtime and request size before retrying. Adjust the timeout in that profile’s detailed settings when needed. Cancel on a chat block stops that request.

For online profiles, Use provider default output length sends no app output limit. Turn it off and set Maximum output tokens to cap a response; provider limits and credit requirements still apply. Local profiles use their own Maximum output tokens setting, initially 8,192. Test the connection and Save in the profile editor.

Inspect AI request details

In the dashboard's AI Model Status panel, select a request row, or focus it and press Enter, to open AI request structure. Review the model, message roles, input types, source filenames and enabled tools.

The details show the most recent prepared provider request, without message text or credentials. Opening the window does not send another AI request. If the request is still being prepared, the details update when its structure becomes available.

Use the request status and its answer or error to check the outcome. The structure window shows prepared inputs; it can still show them when a request fails or is cancelled.

Chat voice input

Voice input uses your installed speech model to turn microphone audio into chat text. AI replies use your configured provider and may incur charges. Automatic sending is on by default after 1.5 seconds of silence following speech.

1. In AI chat, open Mic source, choose a microphone and select Confirm. Install and select a speech model first if Voice input is unavailable.
2. To review drafts before sending, switch off Voice auto send. In AI voice config, turn off Enable voice commands and select Save. With automatic sending on, a pause sends the current draft, including typed text and draft attachments.
3. Select Voice input and wait for the listening status. Dictate your message. The microphone keeps listening after a message is sent; select Stop voice input when finished. Stopping keeps the available draft text.

Open AI voice config to change microphone sensitivity, the silence interval, automatic sending and voice commands, then select Save. Saved changes apply while listening. Sensitivity also affects meeting microphone transcription; system output audio is unaffected.

Voice commands are enabled by default. Say a configured phrase separately, then pause, to delete the last voice segment, clear all input or send the message. Review or edit the phrases in AI voice config; they do not change with the interface language. Turn off Enable voice commands to dictate those phrases as ordinary text.

Live summaries and running memory

The dashboard and Detailed analysis share one live-analysis pipeline. Summary: On summarizes completed transcript chunks; the growing chunk waits until it closes. Settings → Live summary controls chunk duration, the minimum completed chunks per request and previous-summary context.

In AI context settings → Chunk summary, enable previous chunk summaries or running meeting memory when needed. Open Running memory beneath the results to review its overview, key points and evidence ranges. Change the JSON only when you intend to edit the saved memory.

Settings → AI features provides five detail levels for live chunk summaries, selective summaries, chat partial summaries and whole-meeting summaries. Script generation has its own five narration levels; mouse highlights have five density levels.

Analyze and translate speech

Use Summary and Translation in Meeting Control, or open Detailed analysis for transcript, live results and selective results in separate columns. Both windows show the same saved analysis; opening a window does not start a second pipeline.

Translation offers On, Off and Auto. Auto checks each transcript block and translates another language or mixed languages; matching or unknown languages are skipped. Use Change target language in Detailed analysis to choose the output language.

Select transcript text and choose Analyze selected text, or use Analyze time range in Detailed analysis. Summary and Translation control the requested outputs independently. Source shows a result’s input; Update regenerates it, and Cancel stops its pending request.

Review Settings → Live translation for request intervals and preceding context, and AI context settings for each task’s shared content. Online analysis sends enabled inputs to your provider; local analysis uses the selected local model.

Clear selective analysis removes selective results while live work continues. Clear live analysis removes saved results from the chosen chunk onward for every target language and resets chunk-summary memory. Back up results you need before clearing them.

Generate speech and presentation videos

Open Speech generation beside Settings. Task overview links to Text speech generation, Slides speech generation and Files. Generation workflow shows completed steps, required inputs and Results; Generation instructions provides setup details.

Use TTS model management to install a speech model for your device. Enter a narration draft, use Add text file to replace it with a UTF-8 TXT file, or Generate script from meeting to choose transcript, summaries or translations, a target language and detail level.

Select Generate speech from text. Check the language, voice, speed and any required reference audio or transcript, then generate. The draft saves automatically. Speech synthesis uses installed local models; AI-written scripts use your active AI profile.

View model settings lets you inspect another model’s language, voice and reference options without changing the generation model. Those other-model fields are read-only. To generate with another model, return to Choose TTS model in the workspace and reopen generation parameters. Speech speed defaults to 1×, supports 0.5×–2× and preserves pitch.

Speech language may be preset from the draft or the first nonempty slide script. Check the language and voice before generating. If Model information warns that the script language is unsupported, choose a compatible model in TTS model management. Controls vary by model: fixed or automatic language and unsupported voice or reference options are explained in the form.

In Slides speech generation, use Add slides for a PDF, PPT/PPTX, slide images or an image folder. PowerPoint conversion requires LibreOffice. Import numbered slide scripts or use Generate scripts for slides, then review each slide’s narration and target language.

Generate speech from slides before Generate presentation video. Mouse highlights are optional: record or edit them manually, or use Generate mouse movements with AI. Optional fine timing needs Qwen3-ForcedAligner-0.6B and existing slide audio; select the language that matches the narration.

Review video settings and generate the MP4. Completion notices link to View all files; each workflow’s Results offers Open file or Open folder. Hide old results keeps the latest relevant outputs visible. Load meeting folder reopens saved drafts and files; Cancel generation stops active work.

Use only voices and reference material you have permission to use. Review generated narration, timing and highlights before sharing audio or video.

Choose a TTS model

Text to speech (TTS) reads written narration aloud. Install TTS models separately from speech recognition (STT) and the AI model that writes your script.

1. Open Speech generation → TTS model management. Choose a model and the CPU or CUDA installation offered for your device.
2. Review Model information for supported languages, voices, reference requirements, license and estimated download or memory needs, then install. Wait for Installed before selecting it in the workspace.
3. Use a built-in voice for a quick start, or a model with reference-voice support when you have permission to use the recording. CosyVoice also requires the matching reference transcript.
4. Open generation parameters, check Speech language and the available voice or reference inputs, then try a short draft before generating longer narration.

Export a narrated video

Create a reusable explainer, training video or narrated lesson from your slides. The video combines slide images with generated narration and optional timed highlights.

1. Import the slides and review one narration script per slide. PDF, PPT/PPTX and slide images are supported; PowerPoint conversion requires LibreOffice.
2. Generate speech from slides, then review the audio. Add mouse highlights manually, by recording, or with AI; highlights and fine timing are optional.
3. Select Generate presentation video. Choose the result filename, resolution, frame rate, video and audio quality, and pause after each slide. Higher settings increase file size and export time.
4. Generate the video and wait for completion. Use View all files or Open result folder to find the MP4, then play it to check narration, slide order and highlights before sharing.

Write slide scripts

Open Generate scripts for slides and choose Target language and Detail level. Generate all by AI writes scripts for the full deck. For one slide, use Generate by AI and choose its supporting slides. The nearest slides are selected initially; Select all, Deselect all and Restore default change that selection. No supporting slides is also valid.

Edit script text directly in the table, then click Save manual changes. Save those edits before AI generation or Clear text. AI-generated scripts save automatically. Closing with unsaved edits offers Save, Discard changes or Cancel; choose Save when you want to keep the edits.

The script’s Target language controls the written narration independently of the TTS model’s Speech language. Review both before synthesizing speech. View details opens the full script without editing it, and selecting a slide thumbnail opens its zoomable preview.

AI generation asks before replacing existing scripts. Changing or clearing scripts invalidates the current narration, fine timing and video selection, so regenerate the affected speech before exporting a new video. Cancel generation stops the active AI task while keeping previously saved scripts.

Script generation uses the active AI profile. Online models receive the selected slide PDF pages; local vision models receive extracted text and preview images. Review AI slide scripts in AI context settings before including real presentation content.

Voice references

In TTS model management, select the model and CPU or CUDA installation offered for your device. Read its capabilities and license before installing. Wait for Installed, then choose that model in Speech generation. STT models and local chat models are separate installations.

Open Add source files → Add reference audios to import a voice sample. Add reference audio transcript imports a nonempty UTF-8 TXT transcript. To record a sample, choose Audio source, Start recording, then Stop and save. Imported references are copies; removing a copy keeps the original file.

In the generation form, review Reference inputs for the chosen model. Select the audio and, when supported or required, its transcript. Pair files with the same base name before the extension, such as voice.wav and voice.txt, for automatic transcript selection. Choose a differently named transcript explicitly and check its preview against the reference audio.

When dropping a TXT file, choose Narration script to replace the draft, or Reference audio transcript to add reference text. These roles serve different purposes. Review any replacement warning before confirming.

Mouse highlights

Open Generate mouse movements after preparing the slide scripts and narration. Use Edit mouse movements for a slide to add or edit highlights manually, or Generate by AI for that slide. Generate all movements by AI processes the requested slides using your active AI profile.

Record mouse movement opens the recorder without starting it. Choose Start recording for silent recording, or Start recording with audio play to follow the slide narration. Stop recording, review the highlights and Save mouse movements. Play mouse movements opens a separate preview; playback speed changes do not rewrite saved timing.

Mouse styles opens the shared style library. Choose a shape or import an image, adjust its appearance and default size, then Save styles. In the editor, review each highlight’s start and end times, positions, style and size before saving.

Fine timing is optional. Install Qwen3-ForcedAligner-0.6B under Alignment models, generate slide audio, then select the matching language and device. Generate fine timing aligns the final audio with its script; View or Edit lets you review sentence ranges. Regenerating narration or replacing scripts clears their fine timing. Requiring fine timing blocks AI highlights until the requested slides are aligned.

Speech files and results

In Slides speech generation, the Slide script & audio column shows each slide's narration and audio status. Use the script row's Open file or Open folder links for its text file; after narration is generated, the audio row provides separate links for that slide's WAV file.

In each speech workflow, Results lists generated audio and video with Open file and Open folder links. View all files opens Files, which also includes scripts and references. Only generated audio and video are marked Result. Hide old results changes the visible list without deleting files.

Open source folder contains drafts, slides, references and editable timing or highlights. Open result folder contains final audio and video. Keep the entire meeting folder to reopen the workspace later. Load meeting folder changes the meeting in the main window too; it saves the draft and stops speech tasks before switching.

If generation fails, open Generation log and check the selected model, device, required references and free disk space. Cancel generation stops the active task. A completion notice appears after the output is validated and saved; a moving progress bar alone does not confirm success.

Arrange your reading space

In Settings → Other settings, adjust Display size for transcript, summary, chat and translation text, and Interface size for controls and dialogs. Both range from 8 to 14 points; Display size defaults to 11 and Interface size to 10. Fonts offers ten choices per interface language. Changes apply immediately and save automatically; unavailable fonts use the language default.

Use Block title color → Choose color in Other settings to change block headings. The color applies to existing and new blocks. Navigating to a block still highlights its title row so you can find it quickly.

Use Hide or a sidebar toggle to hide a column without deleting its content. Restore it from the sidebar. Fold all and Unfold all control the blocks in each column; Summary and Translation sections also fold independently. Hold Ctrl and scroll over a text pane to zoom it.

Use Analysis links in the transcript to reveal matching live results. Fold or unfold blocks to focus your reading.

Right-click selected text or a block title for Ask AI, Analyze selected text, Ask AI to explain or Ask AI to summarize. The explanation and summary actions attach the selected content as a file and send it to AI chat. With no selection, a block-title action uses the whole block, including folded content.

Choose saved audio quality

Settings → Other settings → Audio quality controls saved meeting WAV audio. Capture settings → Audio quality controls video soundtracks independently. Select Save in Capture settings to apply video choices. Both offer Low, Balanced, High and Very high, with Balanced as the default. Higher quality uses more storage. WAV changes apply when saving next starts or resumes; video changes apply to the next recording. These choices do not change speech recognition quality.

Capture screenshots and visual sources

Audio and visual sources are selected independently. You can capture useful screen content without starting audio recording.

1. Open Screenshot source to select any combination of the whole desktop, available application windows and cameras. Use Preview to check a source, then select Confirm.
2. Select Capture window for a screenshot of each checked source. Select Capture area to drag a rectangle on screen; MeetSimpler minimizes during area selection.
3. For repeated capture, open Capture settings and choose the interval, then select Save. Check your screenshot sources, then select Auto capture. Click the active capture button to stop.
4. Leave Save only when the image changes enabled to reduce repeated automatic screenshots. The default change threshold is 10%; lower values save smaller changes. Automatic capture compares each source with its last saved image. Manual captures always save.
5. Select Capture folder to view saved screenshots under visual_capture_data/screenshots inside the meeting folder, with a separate subfolder for each source.

When a selected window closes or another screenshot source fails, it is removed from the screenshot selection. Automatic capture continues with the remaining sources. If none remain, choose Screenshot source, confirm an available source and start Auto capture again. Screenshot failures do not stop video recording.

Changing screenshot sources while Auto capture is on restarts the capture interval with your new selection. If saving fails, automatic capture stops. Check the meeting folder's availability and free space, then start Auto capture again.

Capturing screenshots saves images locally. To use selected images in AI chat or summaries, add them through Meeting context config.

Record video from selected sources

1. Before recording, open Capture settings to choose resolution, frame rate, video quality and audio quality, then select Save. Video defaults are 720p, 10 fps and Balanced; changes apply to the next recording.
2. Choose Audio sources before starting. The selected microphone and output are mixed into every MP4. Select None in the audio source picker to record without sound.
3. Select Record video to open Video source selection. Check one or more desktop, application window or camera sources. Video choices are independent of screenshot choices.
4. Use Preview beside a source to check it, then select Start recording to record all checked sources. Each source saves to its own MP4 file.
5. Click the active video recording button to stop. Wait for its Saved status before opening the files.
6. Select Capture folder. MP4 files are saved under visual_capture_data/videos inside the meeting folder, with a separate subfolder for each source.

Video recording does not automatically add its frames to AI summaries.

A camera being recorded is unavailable for screenshots or Preview until recording stops. Desktop and window screenshots can continue.

Audio and video source changes apply to the next recording. The current recording keeps the sources selected when it started.

Summarize speech and visuals together

1. Open Meeting context config and add selected screenshots, other images, PDFs or supporting notes.
2. Open Settings → AI context settings and select Customize. Enable Meeting context files for each summary type that should receive those files, then select Save.
3. Use an AI provider and model that support the images or PDFs you include. Generate a partial or final summary, or let configured automatic summaries run.
4. For AI chat, enable Meeting context: On and the context files setting for AI chat. Enable Transcript context: On when you also want the confirmed transcript included.

For example, ask for a summary that connects a slide's key points with the spoken discussion. Check the result against both sources. Selected files go to your configured provider; saving screenshots does not upload them automatically.

Desktop shortcuts and quick selection

Screenshot shortcuts work while another app has focus. Change them in Capture settings if they conflict with another app. Screenshot and video pickers each have their own checkboxes and Preview links.

Name and organize meetings

Open Rename meeting folder. Use Generate AI suffixes for suggestions from the confirmed transcript, or enter a name manually without an AI request. Generating suggestions keeps an active meeting running.

1. If you generated suggestions, select one in your current interface language or English to fill Final folder suffix. With an English interface, suggestions are shown only in English.
2. Edit Current folder name and the optional Final folder suffix. Leave the suffix blank to change only the name; otherwise the two fields are joined with a hyphen. Check Final folder name before confirming.
3. Select Confirm and rename. This stops the active meeting and visual capture, saves the records, and renames and reloads the folder. Select Resume meeting to continue audio capture in the renamed folder; start screenshot or video capture separately when needed.

Retrieve your meeting records

Your records are saved locally in the meeting parent folder chosen during setup. Each meeting has its own timestamped folder. You can reopen that folder in MeetSimpler or read its saved files in Windows File Explorer.

Reopen an earlier meeting

1. If a meeting is running, select Click to stop meeting and wait for saving to finish.
2. Select Load meeting folder in Meeting Control. Browse to your meeting parent folder and choose the individual meeting folder, not the parent folder or a subfolder inside the meeting.
3. Wait for the records to load. Review Live Transcript, then select AI chat to read the latest saved chat session. Use Meeting folder in Meeting Control or AI chat to see the saved audio, transcript and summary files in File Explorer.

Loading a meeting lets you review its saved records. Select Resume meeting only when you want to continue capturing audio in that meeting.

Find saved screenshots and videos

Select Capture folder from the dashboard. Screenshots are under visual_capture_data/screenshots and videos are under visual_capture_data/videos, each organized into source subfolders. Keep the entire visual_capture_data folder with the meeting when backing up your records.

Retrieve a complete audio recording

1. End the meeting and wait for saving to finish, or load the saved meeting you want to export.
2. Select Export full audio in Meeting Control. Choose a destination and a WAV filename. Keep the destination outside audio_segments and use a name different from the original source audio files.
3. Wait for the export to finish, then open the exported WAV file with your audio player. The export contains only audio that was actually saved; periods when audio saving was off cannot be recovered from the transcript.

Read, copy and back up text records

Markdown (.md) files are plain text. Open them with Notepad or a Markdown viewer, then copy the transcript, summary or chat text into your notes or a document. You can read these saved files without generating another AI response.

To keep a restorable backup, wait for recording, AI requests and saving to finish, then copy the entire individual meeting folder, including all files and subfolders, to your backup location. This preserves chat history, audio, analysis, screenshots, videos and meeting state. Use Load meeting folder to reopen the copied meeting later.

For your next meeting

Select Start another meeting to prepare a fresh meeting, confirm the audio sources and then select Start meeting. Use Meeting context config to add an agenda, names or supporting notes, and review the AI context settings before sending that information.

To revisit earlier work, select Load meeting folder and choose that meeting's individual folder inside your parent folder. Review the restored records, or select Resume meeting when you want to continue capturing in the same meeting.

If you already have a recording, use Import audio and follow its file or folder selection and transcription instructions instead of starting live capture.

Common setup problems

For error details, open Settings → Other settings → Account and support → Open system log. Check the relevant entry or STT installation log before retrying. Keep API keys and private meeting content out of support messages.

Still need help? Email meetsimpler@gmail.com with the step you reached and the error message. Keep passwords, API keys and private meeting content out of the message.