How to Use GiliSoft AI Toolkit on Windows

Find a Tool on the Dashboard

1. Choose a category on the dashboard

   Use All to view the collection, or Image, Text, Audio, and Video to narrow the entries.

2. Open the named tool for your source

   Choose Picture to Text for image text, Text to Audio for a script, or Speech to Text for a recording. Use Voice Cloning, AI Face Swap, or a photo tool for those tasks.

3. Check the opened tool's inputs and output controls

   Confirm that you have the intended tool before adding files. Some entries open a separate window with its own language, save location, and processing buttons.

Correct an OCR Result

1. Compare the result with the source image

   Check Recognition result for missing lines, merged columns, wrong characters, and the order of headings or labels.

2. Improve the input or language and recognize again

   Use a sharper, straighter source where possible and choose the correct Language. For a difficult layout, work from a clearer or more focused image rather than assuming OCR recreates a page design automatically.

3. Correct the text where you will use it

   Select Copy content, paste the result into your document, and check names, amounts, dates, and punctuation. Save the reviewed document with the source image still available.

Turn a Recording into Text

1. Open Speech to Text and add the recording

   Choose Speech to Text on the dashboard. In the Audio to Text window, use Add File to import a supported audio or video source.

2. Set Language and Save directory

   Choose the spoken language and review Save directory. Use Source Directory or Custom for the intended destination.

3. Select Extract and review the text

   Use Extract, inspect the returned text and saved result, and compare important passages with the recording. Keep the original for corrections.

Review and Reuse a Transcript

1. Check the text against the recording

   Listen to sections containing names, numbers, abbreviations, or unclear speech and compare those sections with the recognized text.

2. Edit the document for its intended use

   Correct recognition errors and organize notes or paragraphs in the application where the transcript will be used. Keep quotations distinguishable from your own summary.

3. Save and reopen the reviewed text

   Retain both the source recording and the corrected text. If you need timed subtitles, verify a tool's actual timecode and subtitle-export options rather than renaming the transcript as a subtitle file.

Turn a Script into Spoken Audio

1. Open Text to Audio and enter the script

   Choose Text to Audio on the dashboard. In Text to Speech, type into Text content or use Load File for a supported script file.

2. Choose the voice and listen

   Set Language and Speaker, review Speech rate and Volume, then select Preview. Correct wording, punctuation, or pronunciation before generating the complete result.

3. Set the output and Convert

   Choose the offered Output format and Save directory, then select Convert. Open the saved audio in a player and check that it contains the intended script.

Adjust Pace, Pauses, and Background Music

1. Review pace and pauses

   Adjust Speech rate and use Insert silence where the script needs a pause. Preview the affected sentence before changing the full script.

2. Balance voice and optional music

   Use Volume for the speech. If adding Background music, choose the audio and review Background volume so narration remains clear.

3. Convert and check the saved audio

   Listen for rushed names, unnatural gaps, clipped endings, or music that competes with speech. Revise the script or settings, then create and verify a new output.

Open Batch or Long Text to Speech

1. Choose the tab that fits the text task

   In Text to Speech, use Batch Text to Speech for several items or Long Text to Speech for a longer script.

2. Add the content and review that tab's settings

   Use the input controls shown in the selected tab and check its available language, speaker, output, and destination settings. Do not assume the short-text panel's displayed character count is the limit for every tab or license.

3. Test the output before a larger run

   Generate a representative item or section using that tab's controls and check pronunciation and the saved audio. Verify the beginning and end of the remaining outputs after processing.

The Text to Speech screenshot shows the separate batch and long-text entries in the left navigation.

Prepare a Custom Voice

1. Prepare a clear reference recording

   Use a clean, single-speaker recording of your own voice, with little background noise and no competing music. Keep the source available for another attempt.

2. Open Voice Clone to add the reference

   Choose Voice Cloning from the dashboard and select Voice Clone in the opened window. Add the reference and complete the details requested in that form.

3. Review status before generating new speech

   Use Task Center to check the submitted task where applicable. Once a prepared voice is available, select it from List of Speakers and follow the speech-generation task below.

The main voice window is shown in the Voice Clone screenshot. The reference-recording form opens from Voice Clone; the text-entry area is used for the script to be spoken.

Generate Speech with a Prepared Voice

1. Choose a prepared speaker

   Select the intended entry under List of Speakers and use Audition to check the speaker where a sample is available.

2. Enter the new script

   In Text to Voice, enter the text or use the offered Import File control, then choose Text language. Keep the script short for the first test.

3. Set Save directory and Start

   Review Save directory, then select Start when the required input and voice are ready. Listen to the saved audio and check names, pronunciation, pace, and the complete ending.

Swap a Face in a Photo

1. Load the base photo and the face image

   Open AI Face Swap. Use Upload original image for the photo to edit and Upload face image for the face to use. The two inputs have different roles.

2. Choose which face to replace

   Under Select the face to be replaced, select the intended detected face from the original photo. Use Replace beside an input if the source needs changing.

3. Select Start and inspect the saved result

   Choose Start, review the generated photo, and save a new copy using the result controls offered. Check face edges, angle, lighting, skin tones, and retained details outside the replaced area.

Remove a Photo Background

1. Open Remove BG and load the photo

   Choose Remove BG from the dashboard and load the subject image in the cutout tool. Use Replace when another source is needed.

2. Run Cutout and review transparency

   Use Cutout and inspect hair, clothing, and subject edges. For a transparent result, review Keep transparent and turn off a background filling or blur treatment that is not intended.

3. Review Background size and Save

   Check Background size, then use Save with a transparency-capable output setting if transparency is needed. Open the saved image in the receiving app and check the background and edges.

Replace or Blur the Background

1. Start from the cutout and choose the treatment

   After reviewing the subject edges, choose Background blur for a blurred backdrop or Background filling for a replacement background.

2. Adjust the chosen background

   Review the blur slider, or choose Color or Picture under Background filling. Check Background size and the subject's framing.

3. Save and inspect the composition

   Use Save to create a separate result. Check the complete subject, hair and clothing edges, background treatment, and dimensions in the saved image.

Remove an Image Mark or Object

1. Open Image Watermark Removal

   Choose Image Watermark Removal from the dashboard and load the photo. Keep an unchanged original.

2. Select and remove the unwanted area

   Use Smart Pen, Brush, Shape, or Lasso. Review Radius, correct excess selection with Eraser, and select Remove for the chosen region.

3. Inspect at full size and Save Image

   Use 1:1 to inspect the repaired area. Undo and refine a weak result, then use Save Image to save a new copy and check it outside the tool.

Open Photo Repair or Enhancement

1. Choose the photo task from the dashboard

   Use Enlarge for image enlargement, Image Enhancement for a visual improvement, Face Repair for face detail, Colorize Photos for colorization, or Old Photo Repair for an old photograph.

2. Load a copy and review the selected tool's options

   Add the source in that tool and choose the processing options it offers. Start with a representative photo rather than assuming every entry shares the same settings.

3. Process, save, and compare with the original

   Use the tool's own processing and save controls. Inspect text, faces, fine detail, colors, and the complete saved image before using the same settings on more photos.

Find Chat, Translation, and Video Tools

1. Find the exact entry for the next task

   Use All or the relevant category to find AI Chat, Translation, Text to Video, AI Video Watermark Removal, or AI Video Face Swap.

2. Open that tool and add its required source

   Use the selected tool's own text, document, image, or video inputs and review the settings and processing requirements shown there. The photo face-swap controls are not instructions for video face swap.

3. Check and save the result for its intended use

   Review wording and meaning for text, or the relevant frames and sound for video. Save a new output and inspect it in its receiving application before continuing with a larger job.

Choose Where Results Are Saved

1. Review the destination in the active tool

   Check Save directory in the OCR, speech, transcription, or voice tool. Where offered, choose Source Directory to save near the source or Custom for another folder.

2. Choose a clear, writable destination

   Keep originals and results distinguishable and use a folder with enough free space. In photo tools, review the destination offered by Save or Save Image.

3. Open the actual saved file

   After processing, locate the output and open it in a document editor, audio player, or image viewer. A preview or task-list entry is not the same as a checked saved result.

Check Trial and Activation Requirements

1. Identify the installed tool and its trial state

   Check the title and any Register, activation, or restricted-output message. The OCR example includes Register to view all content.

2. Review the requirement for the task you need

   Check that tool's trial limits and any setup or connection prompt. Do not assume every tool has the same limits, one processing allowance, or identical online requirements.

3. Choose the intended license and verify access

   The purchase link in this guide shows $39.95 / 1 Year / 1 PC. Review the plan on the purchase page, then verify the relevant tool's activation and required output before a larger task.

Check the Saved Text, Audio, or Image

1. Check the content, not just whether a task finished

   For text, compare names, numbers, punctuation, and reading order. For audio, listen to the complete narration and ending. For an image, inspect edges, faces, colors, and fine detail at full size.

2. Verify the actual saved output

   Open the saved text, audio, or image in the application where it will be used. Confirm the file, format, dimensions or duration, and destination.

3. Retain a source and a reviewed copy

   Keep the original and the checked result with clear filenames. If a problem remains, revise the source or relevant settings and create another output instead of repeatedly overwriting the only copy.