Files you add to DataReport never leave your phone or computer.
Nothing goes to a serverOpening the file and working on it both happen inside the browser you are using right now. There is no step that puts your file on the internet.
Nothing is keptResults are drawn on your screen and are gone when you close the page. A result file only appears on your own device, when you save it.
There is no accountNo sign-up and no log-in, so there is nowhere that could collect who added which file.
A message you send yourself is the one exception: pressing Report this format in a KakaoTalk tool opens an email, and the report you checked goes out only when you send it.
An export file is usually something you look at once. Opening a page and closing it again, with no install, no sign-in and no permissions, is the safer shape for data this personal.
What is a KakaoTalk chat export?
KakaoTalk can write a conversation out as a plain .txt file (export chat). It is one long flat log of time, sender and message: accurate, but far too long to read as text. These tools turn that file into something you can actually use.
Can I save or share what comes out?
KakaoTalk Chat to CSV writes a CSV file, KakaoTalk Chat to PDF goes through your browser’s Save as PDF, the image tools hand back image files, UUID Generator saves TXT, CSV or JSON, and PDF Spell Checker saves a highlighted PDF and a CSV. The report from KakaoTalk Chat Analyzer is drawn on screen, so take a screenshot of it. Every file is made on your own machine, and nothing is kept on a server once you close the page.
What does it cost?
Nothing. Every tool is free and there is no sign-up.
Photos and images
Are the images I pick uploaded anywhere?
No. This page uses only the browser’s file reading and its canvas, and there is not a single line of code here that puts an image on the network. The compressed file is built on this machine and saved straight to your downloads. Open the page, disconnect from the internet, and the results still come out, so you can check it yourself.
No. This page uses only the browser’s file reading and its canvas, and there is not a single line of code here that puts an image on the network. The resized file is built on this machine. Open the page, disconnect from the internet, and the results still come out, so you can check it yourself.
No. This page uses only the browser’s file reading and its canvas, and there is not a single line of code here that puts an image on the network. The cropped file is built on this machine. Open the page, disconnect from the internet, and the results still come out, so you can check it yourself.
No. This page uses only the browser’s file reading and its canvas, and there is not a single line of code here that puts an image on the network. The converted file is built on this machine. Open the page, disconnect from the internet, and the results still come out, so you can check it yourself.
Does GPS or capture information survive into the result?
No. This tool does not re-save the original file; it builds a new one from the pixels drawn on a canvas. Location, device name and capture time have no route into that file. There is no stripping step, because there is nowhere for the data to sit, which is why the behaviour cannot be switched off. If you are worried about location travelling with a photo you are about to share, one pass through this tool removes it.
No. This tool does not re-save the original file; it builds a new one from the pixels drawn on a canvas. Location, device name and capture time have no route into that file. There is no stripping step, because there is nowhere for the data to sit, which is why the behaviour cannot be switched off.
Thirty per run. Pick more and the first thirty are processed while the screen says the rest are still waiting; nothing is dropped silently. Each file is read up to 30MB. Images are processed one at a time, because expanding several at once overlaps memory and closes the tab on a phone.
Thirty per run. Pick more and the first thirty are processed while the screen says the rest are still waiting; nothing is dropped silently. Each file is read up to 30MB. Every image in a run gets the same width and height, and with the aspect lock on, the side you typed is the reference while each image’s remaining side follows its own ratio.
Thirty per run. Pick more and the first thirty are processed and the screen says the rest are still waiting, so nothing is dropped silently. Each file is read up to 30MB. Images are processed one at a time, because expanding several at once overlaps memory and closes the tab on a phone.
Thirty per run, all saved in the same format. Pick more and the first thirty are processed and the screen says the rest are still waiting, so nothing is dropped silently. Each file is read up to 30MB. Images are processed one at a time, because expanding several at once overlaps memory and closes the tab on a phone.
It reads JPEG, PNG, WebP, AVIF and HEIC, and writes JPEG or WebP. A file it could not read stays in the list with a line saying what to do about it. Before offering a save format it actually builds a one-pixel image in that format and checks that what came back really is that format; it does not guess from the browser name. There is no PNG output, because PNG barely reduces the size of a photograph. Saving a PNG with transparency as JPEG fills the transparent area with white, so pick WebP when transparency has to survive.
It reads JPEG, PNG, WebP, AVIF and HEIC, and writes JPEG or WebP. A file it could not read stays in the list with a line saying what to do about it. Before offering a save format it actually builds a one-pixel image in that format and checks that what came back really is that format; it does not guess from the browser name. Saving a PNG with transparency as JPEG fills the transparent area with white, so pick WebP when transparency has to survive.
It reads JPEG, PNG, WebP, AVIF and HEIC, and writes JPG, PNG or WebP. A file it could not read stays in the list with a line saying what to do about it. Before offering a save format it actually builds a one-pixel image in that format and checks that what came back really is that format; it does not guess from the browser name.
For photos you will look at on a screen, somewhere between 70 and 85 is comfortable: the size drops below half while the change stays hard to see. For photos you will print or edit again, 90 and up. Go below 50 only when you have an attachment limit to meet. Moving the slider re-compresses the files you already picked, so you can compare numbers as you go.
A little, but the quality loss stacks. JPEG discards information every time it saves, so running the same photo through repeatedly makes the mush visible before the size moves much. When a result comes out larger than the original, the saving is shown with a plus sign, and that number is not hidden.
If size is the goal, reducing dimensions beats reducing quality. Taking a 4000-pixel photo down to 1600 pixels cuts the pixel count to under a sixth. On this page you only set a long-edge limit, and the aspect ratio is kept. Photos already smaller than the limit are left alone, so nothing is stretched up.
The word on the button tells you exactly what is happening. Share opens the share sheet, where Save to Photos puts it in the gallery. Save opens a location picker so you choose the folder. Download drops the file into the browser’s downloads folder. The Photos app does not read the downloads folder, so a file that arrived that way goes into the gallery once you move it there from the Files app.
The word on the button tells you exactly what is happening. “Save to Photos” opens the share sheet, where you can put it straight into the gallery. “Choose where to save” opens a location picker so you choose the folder. “Download” drops the file into the browser’s downloads folder. The Photos app does not read the downloads folder, so a file that arrived that way goes into the gallery once you move it there from the Files app.
That means the result came out larger than the original. Phone camera photos and files that have already been compressed grow when re-saved at the same resolution. When that happens this tool hands back the original rather than the larger file, because you came to reduce the size, and it does not quietly lower the quality and try again. The reliable move is to set the goal to Email / chat attachment: a long edge of 1600 cuts the pixel count sharply, and the size follows. When the original is handed back, its capture metadata is still the original’s.
Choose Target file size under Goal and the input appears; it stays hidden while another goal is selected. Quality is adjusted first to bring the file under the target. If the quality floor is reached and the file is still over, the long edge steps down through 2560 → 2048 → 1600 → 1280 → 1024 → 800 → 640 → 512 → 400 → 320 and it tries again. Among the results that met the target, the best quality is kept. The result card says what it finally used, for example 500KB target met at quality 72%, with the resolution shown beside the original’s. If the bottom of the ladder still does not reach the target, the card says short of target and hands you the smallest it managed. It does not present a miss as a hit.
If I do not like the result, is my original changed?
No. This tool reads the files you pick and never writes to them. The result is a new file with -compressed in its name, and even that does not touch the disk until you download it. If you do not like it, change the quality or the format and run it again, or clear everything with Start over.
No. This tool reads the files you pick and never writes to them. The result is a new file with -resized in its name, and even that does not touch the disk until you save it. If you do not like it, change the numbers and run it again, or clear everything with Start over.
No. This tool reads the files you pick and never writes to them. The result is a new file with a different extension, and even that does not touch the disk until you save it. If you do not like it, pick another format and run it again, or clear everything with Start over.
Leave Lock aspect ratio on and type only one side. The other is calculated from the original ratio and filled in; where the calculation lands on a fraction it is rounded to the nearest whole pixel, so the error against the original ratio is at most one pixel. To force exact width and height regardless of the ratio, turn the lock off and type both, and the image will stretch or squash. If what you need is to cut part of the picture away, Image Cropper is the right tool, not this.
The length is converted to inches and multiplied by the DPI. One inch is 2.54cm, so 10cm at 300 DPI is 10 ÷ 2.54 × 300 = 1181 pixels. That result is shown right beside the field as → 1181 px, and the file that comes out has exactly that pixel value. Rounding happens in exactly one place, so the number on screen and the number in the file cannot drift apart.
For a print shop or a photo print, 300. For a document coming out of a home printer, 150 is enough. For an image you will only look at on a screen, DPI means nothing, so set the unit to pixels and type the pixel value you want. Raising the DPI raises the pixel count for the same centimetres, so the file gets bigger.
You can, but it will not get sharper. Nothing new appears; existing pixels are stretched to fill, so edges soften when you zoom in. Use it when the destination demands a minimum pixel count. This tool does not block enlarging; instead it writes the original resolution and the result resolution side by side in the list so you can see exactly what changed.
What is the difference between cropping and resizing?
Cropping throws away the outside of the picture and keeps the part you marked. Resizing keeps the whole picture and changes only the pixel count. If a person is off to one side and you want them centred, that is cropping; if you need the whole picture but the file is too big, that is resizing. To do both, crop here first and put the result into Image Resizer.
Yes. Drag inside the box with a finger to move it and drag the corner handles to resize; the handles are sized to be pressable with a fingertip. When that is too coarse, typing the width, height and offsets is exact. The two ways mean the same values, so mixing them never puts anything out of step.
Press a ratio button and the box locks to it. After that the aspect stays fixed however you drag the corners, and typing one side into the number fields brings the other along at that ratio. Press Free to unlock and move each corner on its own. The pixel size shown after cropping is the size the file will actually have. Rounding happens in exactly one place, so the screen and the file cannot drift apart.
Can I crop several images to the same region at once?
Yes. Up to 30 at a time, and the crop region is remembered as a proportion: ask for the middle half and each photo gives you its own middle half even at different sizes. When the originals differ in size the resulting pixel dimensions differ too, so each image’s actual values are written on its own line in the result list. To force every image to exactly the same pixels, crop here and then run the results through Image Resizer.
The quality of the part you keep is untouched. The original pixels are taken from where they are, with no stretching or shrinking arithmetic involved. Saving does re-encode as JPEG or WebP, so pushing the quality value far down costs something at that step. The default of 90% is a range where the difference is hard to see.
No. This tool reads the files you pick and never writes to them. The result is a new file with -cropped in its name, and even that does not touch the disk until you save it. If you do not like it, move the box again or clear everything with Start over.
The file gets much smaller. In exchange, transparent areas are filled with white, and once it is JPG that loss cannot be undone. For a photograph, JPG is right; for a logo, lettering or a shape with hard edges, leaving it as PNG is better, because those turn blotchy along the edges as JPG.
PNG saves without loss, so on a photograph, where the pixels scatter finely, it comes out far larger than JPG. Measured in this tool, several times larger is common. That is not an error but the normal result of that combination, and the screen says plainly that it grew. Unless you are converting because you need transparency, there is no reason to keep a photograph as PNG.
For images going onto a web page, WebP is usually the smallest: smaller than JPG at the same visual quality, and it can hold transparency. For a file you are sending to someone else, pick JPG; it opens with whatever they have.
That is JPG: the format cannot hold transparency, so transparent areas are filled with white. Pick PNG or WebP when transparency has to survive. This tool does not hide that; the result line says which format the file went out as.
No. This tool changes only the format; the pixel dimensions stay exactly as they were. To change dimensions, use Image Resizer; to hit a target file size, use Image Compressor.
Drop the images onto the page or pick them with Choose files; several at once is fine. Moving the quality slider re-compresses the files you already picked and shows the result. Choose JPEG when the file has to open anywhere, WebP when size matters more. Set a long-edge limit and only photos above it are scaled down, keeping their aspect ratio; Keep original leaves dimensions alone. Download image by image from the result list or press Download all. Before and after bytes and the percentage saved appear next to the original and the result, and pressing either picture compares them full width. Four settings decide the result: quality, format, long edge and how many images. Change any one and the files you picked are re-run, and a result larger than the original shows its saving with a plus sign.
Why will a phone photo not shrink in Image Compressor?
A photo that is already optimised will not shrink at the same resolution. The JPEG a modern phone writes has already been through the camera app’s compression, so saving it again at the same dimensions only costs quality while the bytes stay put or go up. Resolution is what moves the number: a 4284×5712 photo is 24 megapixels, and bringing the long edge down to 1600 cuts the pixel count to about one thirteenth, with the file size following. Email / chat attachment (long edge 1600) is the default and looks all but identical on a phone or monitor. Web upload (long edge 2048) is for places people zoom into, like a blog post or product page. Target file size meets a fixed limit by adjusting quality first and stepping the long edge down if needed, then says which quality and resolution it used. Keep resolution (quality only) is for when the dimensions must stay, and barely moves an already-optimised photo. Each result card shows the original resolution and the result resolution side by side.
How do the quality ranges work in Image Compressor?
The quality number is not a percentage of the original; it says how much information is kept when saving. That is why the same 80 percent shrinks a smooth landscape far more than a screenshot full of text and lines. 90 and up is hard to tell apart by eye and suits photos for print or editing, with 20 to 40 percent off the size. 70 to 85 is the default for anything on a screen, such as blog posts, messengers and attachments, and the file usually drops below half. 50 to 65 is for when the size has to come down a lot, and blotching starts to show in smooth areas such as sky and skin. Under 50 is only for meeting an attachment limit, and the mush is obvious when you zoom in. You rarely need to set the number yourself: picking a goal sets quality and long edge together, and the slider sits folded inside Advanced.
Picking a goal sets quality and long edge for you. Change one by hand under Advanced and the goal moves to Keep resolution, with your value used as given. The before and after size in the result list updates immediately, so you can find the point you want by eye.
What is the difference between JPEG and WebP in Image Compressor?
JPEG opens everywhere and is strong on photographs where colour runs continuously; it cannot hold transparency and fills it with white. WebP is 20 to 30 percent smaller than JPEG at the same quality, for the web and for reclaiming your own storage. For a file whose recipient you do not control, such as an email attachment or a form submission, leave it on JPEG; for images going onto a web page, WebP is smaller at the same quality.
How does Image Compressor keep capture data out of the result?
A photo taken on a phone carries more than the picture: the coordinates where it was taken, the device name and the capture time, so one picture taken at home carries your home address as coordinates. This tool does not repackage the original file; it builds a new file from the pixels, so capture metadata has no route into the result. There is no stripping step that runs, simply nowhere for it to go, which is why this is always on. Orientation is read by the browser and applied, so results do not come out sideways or flipped. A photo you have already posted somewhere cannot be walked back with this tool, so metadata has to come off before it goes out.
There is no upload path in this tool. Every calculation happens inside this browser, which is why compression finishes with the network off. There are no accounts and no database, so closing the tab clears it from memory, and the result file touches your disk only when you press download. The KakaoTalk tools here, KakaoTalk Chat Analyzer, KakaoTalk Chat to PDF and KakaoTalk Chat to CSV, are built on the same rule and keep the file inside your browser.
It changes an image’s width and height right inside this browser, in pixels or in cm and mm, with the picture you put in and the picture that comes out side by side.
Drop the images onto the page or pick them with Choose files; several at once is fine. Choose a unit: pixels for anything on a screen, cm or mm for print, which brings up the DPI field. Type the width and height; with Lock aspect ratio on, giving one side fills the other from the original. If you typed cm or mm, → N px appears beside the field, and that number is the pixel value the file will actually have. Save image by image from the result list or press Download all; original and result appear side by side, and pressing either compares them full width.
How does Image Resizer turn cm and mm into pixels?
The length is converted to inches and multiplied by the DPI, one inch being 2.54cm. The result appears as → N px beside the field, and the file that comes out has exactly that pixel value. The common print sizes are in the Common sizes buttons as pixels at 300 DPI, so you can press instead of converting centimetres by hand. Which DPI to pick is covered in the question What DPI should I use?
What happens in Image Resizer when the aspect lock is off?
To force both sides regardless of the ratio, turn the lock off and type both values. The image then stretches or squashes. If what you actually need is to cut part of it away, Image Cropper is the right tool rather than resizing.
Up to 30 at a time, all with the same settings. With the aspect lock on, the side you typed is the reference and each image’s remaining side follows its own ratio; to force every image to exactly the same width and height, turn the lock off and type both values. Images are processed one at a time, because expanding several at once overlaps memory and closes the tab on a phone.
What is the quality setting under Advanced in Image Resizer for?
Resizing re-saves the file anyway, so the quality value affects the result size too. The default of 90% is a range where the difference is hard to see.
How does Image Resizer keep capture data out of the result?
GPS coordinates and the rest of the capture metadata do not survive into the result. The original file is not repackaged; a new file is built from the pixels, which is why this is always on and cannot be switched off.
This page does not take a size limit. Smaller dimensions do bring the size down, but to meet a fixed limit such as an attachment cap, use Image Compressor.
Which sizes are in the Image Resizer common size buttons?
For anything you will look at on a screen, pixels are the exact way to say it, and the values people reach for most are in the Common sizes buttons; one press fills both fields. On screen: Full screen / slides 1920×1080. Fits a full-screen background or a presentation slide. Square post 1080×1080. The size square social posts use. Into a document 800×600. A comfortable size for dropping into a document. For print: 4×6 inch print 1181×1772. The most common photo print size, 10×15cm at 300 DPI. A4 document 2480×3508. For filling an A4 page, 21×29.7cm at 300 DPI. ID photo 413×531. 3.5×4.5cm at 300 DPI. Check the size the place you are submitting to asks for, and enter that. If you are not sure which value is right, use the largest size the destination accepts. Going above it only makes the file heavy, because the other side scales it down again anyway.
It cuts just the part you want to keep out of a photo. Drag the box or pick a ratio, and the picture you put in and the picture that comes out are shown side by side.
Which files does Image Cropper take, and how many?
JPEG, PNG, WebP, AVIF and HEIC files, up to 30 images at a time and 30MB each. The cropping finishes inside this browser and nothing is sent to a server.
Why does Image Cropper leave no location data in the result?
GPS coordinates and the rest of the capture metadata do not survive into the result. The original file is not repackaged; a new file is built from the pixels, which is why this is always on and cannot be switched off.
Drop the images onto the page or pick them with Choose files; several at once is fine. A box appears over the picture marking what will be kept. Drag inside it to move it and drag a corner to resize, and a finger does the same. For a fixed ratio, press a ratio button: the box locks to it and stays locked however you drag the corners. To hit exact pixel values, type them into the width, height and offset fields; the box and the numbers always mean the same thing. Save image by image from the result list or all at once. Original and result appear side by side, and pressing either compares them full width.
Does Image Cropper save exactly the area I set on screen?
If the region you set on screen and the file you get did not match, the crop would have to be done again from scratch. So this tool was built while checking, pixel by pixel, that what comes out is the content of the rectangle the box marked. The width and height on screen are the width and height of the saved file, and the cropped picture is that coordinate region of the original.
It changes an image’s format right inside this browser. Pick JPG, PNG or WebP and it is saved as that, with the picture you put in and the picture that comes out side by side.
Which files does Image Format Converter take, and how many?
JPEG, PNG, WebP, AVIF and HEIC files, up to 30 images at a time and 30MB each. The conversion finishes inside this browser and nothing is sent to a server.
Why does Image Format Converter leave no location data in the result?
GPS coordinates and the rest of the capture metadata do not survive into the result. The original file is not repackaged; a new file is built from the pixels, which is why this is always on and cannot be switched off.
Drop the images onto the page or pick them with Choose files; several at once is fine. Choose JPG, PNG or WebP under Save as. With JPG or WebP you can set the quality; PNG has no quality value, so that control disappears. Save image by image from the result list or all at once; file names stay as they are and only the extension changes. An image whose file got bigger is labelled larger. That is not a failure, because that combination genuinely grows, and the line says exactly what changed and by how much.
Why does Image Format Converter keep the chosen format even when the file grows?
What separates tools here is what they hand you when the file grows. Returning the original because it grew leaves you with a file you believe is PNG that is in fact still JPG. This tool saves in the format you picked even when that is larger, and it never quietly swaps the format on you.
What does the result list in Image Format Converter show?
For every image the result list writes the original format, the result format and the size change. Where some grew and some shrank, each line carries its own numbers.
The v7 values from UUID Generator are not v7 in name only; they are built to the v7 layout. The leading 48 bits carry the millisecond you pressed Generate, and a batch of up to 10,000 stays in ascending order. The same batch leaves as a TXT, CSV or JSON file.
Which versions and shapes can UUID Generator produce?
Version 7 by default, plus v4, v1, v5 and NIL. v5 is built from a namespace (DNS, URL, OID or X500) and a name. Up to 10,000 at once, with hyphens, lowercase or uppercase, braces, URN, quotes and line breaks set by you, then copied or downloaded as TXT, CSV or JSON.
Every UUID is produced inside this browser, and no value is sent to a server. The SHA-1 step that v5 needs runs in the page’s own code too, so v5 works wherever the page is opened.
Every word it finds comes back with the page it sits on and a shrunken render of that page with the word boxed on it, so you can tell one occurrence of a typo from the next without hunting, and the same list leaves as a highlighted copy of your PDF and as a CSV.
You choose between the en-US and en-GB English dictionaries. Words written only in capitals and words containing digits are skipped from the start, and you can switch either setting off. The check runs entirely inside your browser and the PDF is never uploaded.
No. This page uses nothing but the browser’s file reading, and there is not a single line of code here that puts chat contents on the network. Open the page, disconnect from the internet, and the analysis still runs to the end, so you can check it yourself.
How do I see how much was said in a KakaoTalk chat?
In KakaoTalk, use the chat room menu, then Chat history, then Export chat history to save a .txt file, and pick that file on this page. Message counts and character counts per person, and each person’s share of the whole, come back as a ranking table.
Yes. Share of messages and share of characters are calculated separately. They often differ, because someone who sends fewer messages may write much longer ones. Average message length is shown beside them.
Yes. Android, iPhone and PC KakaoTalk write the export differently, and this tool tells the shapes apart on its own, including files exported with the English interface. The format it recognised is named on the format line of the result, so you can see straight away whether it read the file the way you expected.
Yes. Android, iPhone and PC write the export differently, and PC alone varies the time notation between a Korean 12-hour form, 15:24 and 3:24 PM. This tool tells those shapes apart on its own, including files exported with the English interface. A ZIP or an emailed .eml goes in as it is, and the chat file inside is found and read.
Yes. A group export is read as it is. More people simply means a longer ranking table, with messages, characters, share and average reply time computed per person so they can be compared. A real group export was checked against hand-counted figures, and the per-person distribution, hour of day, weekday and attachment tallies all matched. Someone who joined or left partway through is counted only from what the log still holds, so a short stay shows up as a correspondingly small share.
Another tool would not open my file. Will this one?
Add it and you will know at once, and the screen also tells you whether the result can be trusted. This tool reads the head of the file to decide the format. When no format fits, it says the format was not recognised instead of showing wrong numbers. A file whose format was recognised but from which no message could be read is treated as a failure too, because a screen reading 0 messages and 0 participants must never look like a successful analysis.
Add it and you will know at once. This tool reads the head of the file to decide the format; when no format fits, it says the format was not recognised instead of building an empty document. A file whose format was recognised but from which no message could be read is treated as a failure too, so that you are never handed a PDF with nothing in it. When it does read, the message count and A4 pages appear immediately, so you can confirm it before printing.
Drop the ZIP in as it is and the chat .txt inside is found and read. An .eml that KakaoTalk mailed to you is read the same way, with the attached .txt pulled out. For a password-protected ZIP, unzip it yourself and add the .txt.
Up to 100MB or 1,000,000 lines. The file is read in chunks rather than held in memory all at once, with only the statistics accumulating, so long conversations hold up. Beyond that range you get a notice rather than an error, and re-exporting in shorter date ranges from KakaoTalk covers the rest.
It averages the time between the other person’s message and your answer. Gaps longer than six hours are treated as sleep and left out of the average. Breaking a silence of three hours or more is counted separately as opening the conversation.
This page draws the statistics on screen, so take a screenshot of what you want to keep. To keep the conversation itself as a table, put the same file into KakaoTalk Chat to CSV and download a CSV with six columns: date, time, weekday, sender, message and type. For a printable document, KakaoTalk Chat to PDF does that. Neither stores anything on a server, so closing the page leaves nothing behind.
It says some lines could not be read. How do I report that?
Open the lines not read notice in the result and press Report this format to compose a draft. The draft carries the page address, the recognised format, the number of unread lines, and those lines. Names and message text are masked automatically. You can edit or delete the draft before sending. Nothing is transmitted in the process; the contents leave your browser only when you copy the draft and send it yourself.
First make the .txt in KakaoTalk: Settings, then Export chat history. On iPhone an .eml that arrives by mail, or the .zip, works just as well. Add that file with “Choose a file” on this page, press “Open print dialog”, change “Destination” (or “Printer”) to “Save as PDF”, and save. The tool tells apart the export formats that differ by device and by interface language, so files another tool refused are read here.
No. This page uses nothing but the browser’s file reading, and there is not a single line of code here that puts chat contents on the network. The PDF is produced on this machine by the browser’s own “Save as PDF”. Open the page, disconnect, and it still runs to the end, so you can check it yourself.
Why a print dialog instead of a “Save PDF” button?
Because the browser’s print function does the job. Change “Destination” to “Save as PDF” in that window and you get a file. The reason for this design is fonts: for a web page to build a PDF itself it must download and embed an entire Korean font, and on a long conversation the browser stalls doing it. Printing uses the fonts already installed, so that problem does not arise.
Does the conversation come out exactly as written?
Not one character is changed. Spacing, line breaks, emoji and special characters are carried across as they are. Conversions that route through a word processor insert spaces that were never there between a number and the unit that follows it, or drop line breaks; this tool does no such processing. It was checked against a real user’s chat file all the way through printing: whether blank lines and indentation inside multi-line messages survive, whether the (continued) header repeats when a date runs onto the next page, and whether every participant name appears, all compared by eye in the print preview.
A4 portrait, 15mm margins on all four sides, with a page number on every page. A header goes in when the date changes, and when the same date runs onto the next page that header repeats with a (continued) marker.
Reading a file and laying it out for print have different limits. Reading goes to 100MB or 1,000,000 lines; printing has to expand every page at once, so it is lower. What sets the printing limit is the number of A4 pages, not the message count, because the same number of messages takes far more pages when people write in long multi-line blocks. The recommendation is 1,000 pages and the maximum is 2,000; add a file and the tool tells you the messages and A4 pages straight away. At real-world density that is roughly 7,200 and 14,500 messages. Above the recommendation the print window takes a long while to appear, so you are warned first; above the maximum you get a notice rather than an error. Splitting by date range and saving more than once covers the whole thing.
Can I save it with the participants’ names hidden?
Yes. Switch on “Mask participant names” and senders become Participant 1, Participant 2, with the same person keeping the same number throughout, so the thread still reads. Only the sender name changes; the messages themselves stay exactly as written. Names or phone numbers written inside a message are not hunted down and masked, so if there are any, search for them with “Keyword” before you save.
Can I save only one person, or only part of the period?
Yes. Group exports are read as they are and the participant list fills itself in. You can narrow by date range, by participant and by keyword, and the number of matching messages updates as you go. Use it to pull one person’s messages out of a group chat, or to save a single month.
We will only state what the tool does. It carries the original across without changing a character, and the chat file is neither transmitted nor stored. What weight the resulting document carries in a given procedure is not something this tool can judge, so check with the relevant body or your lawyer before submitting.
Is a KakaoTalk conversation admissible as evidence?
Whether something is admitted is decided case by case by a court, so this tool cannot answer that. What we can state is what the tool does: it carries the contents of KakaoTalk’s own exported file across without changing a character, and with “Evidence markings” switched on, the cover page records the original file name, its size, its SHA-256 and how many of the total messages were printed. A SHA-256 changes if the file changes, which is what makes it useful for confirming later that it is the same file. Whether and how to submit it is a matter for your lawyer.
Double-click it; that is the whole procedure. The file is written with a UTF-8 byte order mark so that non-Latin text does not break, which means no import wizard and no settings. Google Sheets, Numbers and Hancell open the same file as it is.
Do system lines like “joined” and “left” become rows?
No. System notices such as “X joined”, “X left” and “X was invited”, and the export preamble at the top of the file (“KakaoTalk chat with X”, “Saved on: …”), were not sent by a person, so they are not turned into rows. The table holds only the messages that were actually exchanged.
That stays. It is not a system notice but a real message with a definite sender. Sender and time are filled in as usual and the type column says it was deleted, so you can filter on that column when you need to count when and by whom messages were removed.
They are classified in the type column as photo, sticker, video, voice message, file, map, contact or deleted message, eight kinds in all. Ordinary messages leave the cell blank, so keeping only the blank rows leaves exactly the text people wrote.
No. Messages containing line breaks, commas or double quotes are all wrapped the standard way (RFC 4180), so each stays inside one cell. In Excel you see the line wrap inside that cell.
This tool makes CSV only. Producing .xlsx would mean loading a spreadsheet library into the page, which adds an outside request and clouds the one thing this tool promises: nothing is sent anywhere. Excel opens a CSV on a double click, and Save As from there gives you an .xlsx.
Add a KakaoTalk chat export and it turns the file into one report: who said how much, when the two of you actually talked, and how fast the replies came back.
An export is a flat log of time, sender and message, over and over, far too long to count by eye. This tool makes one pass over the file and counts by four rules. First, messages and characters are counted separately, because some people send few messages but write long ones, and you need both numbers before who talked more means anything. Second, a multi-line message counts as one: everything up to the next timestamp is one message, so the total is not inflated. Third, photos and stickers are kept out of the character count, because an export writes a photo as the word photo and the person who sent pictures would otherwise look talkative. Fourth, only countable words go into the word ranking; single characters, filler words and attachment markers are skipped.
How are the share of the conversation and the participant ranking shown?
The messages per person section draws each participant’s share as a single bar, with the ranking table beside it. The ranking goes by message count, and the same table carries share of characters, average message length, average reply time and how often each person opened the conversation. A group chat simply makes the table longer; in a two-person chat you can see at a glance whether the split sits near half and half or leans one way. Messages and characters are counted separately so that the person who sends many short lines and the person who writes long ones stop looking alike, and reply gaps are capped so that sleeping hours do not land in the average.
The same app writes the export differently on each device. Android and iPhone put the date, a comma, the name and the message on one line, but write the date in different shapes, and iPhone uses both 12-hour and 24-hour times. PC puts the name and time in square brackets under a date separator line, for example [Name] [3:24 PM] message. This tool reads the head of the file and works out which shape it is on its own, and exports made with the English interface go through the same path.
A chat export carries the other person’s name, every sentence you both wrote, and the timestamps. This tool has no upload path at all: the browser is what reads the file, and every calculation happens inside it, which is why the report finishes with the network off. There are no accounts and no database, so closing the tab clears it from memory, and the analysis runs on its own thread so the page stays responsive while a large file is read.
How do I keep an analysed chat as a table or a printable document?
If you want the conversation itself as a table rather than numbers, put the same file into KakaoTalk Chat to CSV; it comes back as a CSV that opens straight in a spreadsheet. For a printable A4 document, use KakaoTalk Chat to PDF. There, picking a date jumps the preview to that date without changing the number of messages to print, and the word you searched for is marked in the preview so you can check before you save.
How does the chat analyzer show lines it could not read?
Every figure in the report is a count, so what fell outside the count has to be visible too. If some lines did not match the format, the result shows how many under lines not read, and you can open them and look. That display exists on screen only; it is neither stored nor sent anywhere.
How is keeping a chat as a record different from analysing it?
The analysis report is aggregate numbers and does not carry the conversation itself. If you need the conversation as a record, put the same file into KakaoTalk Chat to PDF and switch on evidence markings. The cover page then carries the original file name, its size, its SHA-256 and how many of the total messages were printed.
It turns a KakaoTalk chat export into an A4 PDF. You can filter by date range, participant and keyword, and not one character of the conversation is changed.
Which files does KakaoTalk Chat to PDF take, and how many pages can it print?
It takes .txt, .zip and .eml files and reads up to 100MB or 1,000,000 lines. One print run is recommended up to 1,000 A4 pages and allowed up to 2,000. The page count follows the conversation; at real-world density that is around 7,200 and 14,500 messages. The file is read inside this browser and the PDF is made on this machine.
How is Narrow it down set when KakaoTalk Chat to PDF opens a file?
Everything starts selected: the whole date range and every participant. Clear only what you want left out. If some lines did not match the format, you are told how many before you save. A keyword you type is marked in the print preview, and that marking is on screen only; it does not appear in the printed document.
How is Jump to date different from the date filter in KakaoTalk Chat to PDF?
Jump to date only scrolls the print preview to that date. It is separate from the date filter, so the number of messages to print does not change. To print less, use From and To under Narrow it down.
Why export a KakaoTalk chat as .txt instead of taking screenshots?
Screenshotting the app gives you dozens of images you cannot search. KakaoTalk’s own Export chat history writes a .txt where the text stays text, and KakaoTalk Chat to PDF lays that file out as an A4 document. With no file yet, See an example on the page shows the shape of the result.
Does the PDF converter read an .eml from iPhone mail or a .txt from the Files app?
Yes. An .eml that arrived by mail and a .txt saved in the Files app are both read as they are. To keep photos and video from an iPhone export as well, choose Save all messages as a document, store it on the phone and move it out through the share sheet. Saving the mail itself as .eml works too, and so does the .txt from inside the zip.
How should the print window be set for KakaoTalk Chat to PDF?
Open print dialog brings up your browser’s print window. Change Destination or Printer to Save as PDF and it saves as a file; pick your printer in the same place and it comes out on A4 paper. File or paper, only that one control differs and the document is the same. Paper is already set to A4 portrait. Leave Scale at Default, because shrinking it by hand throws the page breaks off. Clear Headers and footers, because page numbers are already inside the document and would print twice. This tool does not build the PDF itself; it uses your browser’s save.
What if I cannot find the exported KakaoTalk chat file?
Mail it to yourself instead. Send by mail in KakaoTalk gives you an .eml, and that .eml goes straight into KakaoTalk Chat to PDF; the tool pulls the chat file out of it. A .zip you received works the same way.
How do people use a KakaoTalk chat PDF for backup or sending?
Once it is a PDF, backing the conversation up or sending it to someone is one file instead of dozens of screenshots, and you can search inside it for a word. People also use it before changing phones or leaving a chat room. What the other person wrote goes in as well, so to keep only one person’s messages, select just that person under Participants before saving. For statistics such as message counts, shares and time of day, use KakaoTalk Chat Analyzer; for a table to sort and filter in a spreadsheet, use KakaoTalk Chat to CSV. All three take the same file.
What do the evidence markings in KakaoTalk Chat to PDF record?
Screenshots of a conversation cut the context on either side and pile up into dozens of images. The .txt KakaoTalk exports keeps the conversation in order with the text intact, so laying that file out as a document keeps the thread unbroken. Switch on Evidence markings and the cover page carries the original file name, its size and its SHA-256, together with how many of the messages were printed. A SHA-256 changes if a single character of the file changes, which is what makes it useful for checking later that the file is the same one. Even when you filtered by date or participant, the cover still says how many messages the whole file held.
Does masking names in the PDF converter change the original file’s hash?
No. Mask participant names works together with Evidence markings. The SHA-256 is computed from the bytes of the file you added, so masking names in the document does not change it, and you can hand over a masked document while the original file still checks out. What the hash confirms is that the original file is the same one, not that the printed document looks like it.
Can KakaoTalk Chat to PDF print on paper instead of saving a file?
Yes. Press Open print dialog and choose your printer instead of Save as PDF; there is no need to make a PDF and open it again. The document is laid out to A4 from the start, so nothing is cut off on paper, and paper size, margins and page numbers are already set in the print window.
Where does the PDF converter process my chat file?
A chat export carries the other person’s name, every sentence you both wrote, and the timestamps, so there is no reason to put it on someone else’s server just to convert it. This tool was built without an upload path at all. The browser is what reads the file and there is no code that puts its contents on the network. The PDF is made on this machine too, by the browser’s own print function. There are no accounts and no database, so closing the tab clears it from memory, and the conversion still runs to the end after you turn the internet off.
It turns a KakaoTalk chat export into a CSV that opens straight in a spreadsheet, with the conversation split into six columns: date, time, weekday, sender, message and type.
Why does KakaoTalk Chat to CSV need an exported .txt first?
While a conversation lives inside the app there is no way to count it or filter it. KakaoTalk’s own Export chat history writes a .txt where the text stays text, and KakaoTalk Chat to CSV turns that file into a table. No file yet? See an example on the page shows the shape of the table first.
Why does a KakaoTalk Chat to CSV file open with real columns?
An export is a flat log with time, sender and message stuck together on one line, so pasting it into a spreadsheet drops the whole thing into a single cell. KakaoTalk Chat to CSV takes that line apart and writes it back as a CSV with real columns, which is what makes sorting, filtering and pivoting work straight away. A UTF-8 byte order mark at the head of the file keeps non-Latin text from breaking in Excel, and messages containing line breaks, commas or double quotes are wrapped the standard way (RFC 4180) so the columns never shift. Cells that begin with =, +, - or @ are written as fixed text so Excel does not try to run them as formulas.
What does each column in KakaoTalk Chat to CSV hold?
The table is always six columns: date, time, weekday, sender, message and type. The first row is the column names and messages start on the second. The date is exactly as written in the export and no time zone is applied, so messages near midnight do not shift a day. Time is 24-hour, and because KakaoTalk exports carry no seconds none are invented. Weekday is worked out from the date, and sender is the name as shown in the chat room. A multi-line message keeps its line breaks inside a single cell and nothing about the text is changed. Type is filled only for attachments that have no text body. System lines such as joined, left and invited, and the export preamble, are not turned into rows, while “This message was deleted” is a real message with a sender, so it stays as a row and is marked in the type column.
What can I do with a KakaoTalk Chat to CSV file once it is open?
Double-click the downloaded kakaotalk-chat-DATE.csv and Excel opens it directly, with no import wizard and no delimiter to set. Google Sheets takes it through File, then Import, and Numbers opens it if you drag it in. People usually filter on the sender column to keep one person’s messages, sort by weekday and time to see when they actually talked, and keep only the rows where type is blank to count what was really written rather than photos and stickers.
How far does KakaoTalk Chat to CSV go with a very long chat?
It takes .txt, .zip and .eml files and reads up to 100MB or 1,000,000 lines (a multi-line message counts as several lines). It builds up to 200,000 rows, one message per row. Past the row limit it does not error; it keeps the first rows, says the table was truncated, and lets you download it. If no message could be read at all, it builds no empty file and explains why instead. For a conversation beyond that range, re-export from KakaoTalk in shorter date ranges, six months at a time for example, and each file fits.
A chat export carries the other person’s name, every sentence you both wrote and the timestamps, so there is no reason to put it on someone else’s server just to make a table of it. This tool has no upload path: the browser is what reads the file, and the CSV is made on this machine too and saved by the browser’s own download. There are no accounts and no database, so closing the tab clears it from memory, and the conversion still runs to the end after you turn off the internet.
Which tool do I use for chat statistics or a printable document?
For statistics such as message counts, shares and time of day, use KakaoTalk Chat Analyzer. For a printable A4 document, use KakaoTalk Chat to PDF. There, picking a date jumps the preview to that date without changing the number of messages to print, and the word you searched for is marked in the preview (it does not appear in print). All three take the same file, so the .txt you exported once is all you need.
How does KakaoTalk Chat to CSV show lines it could not read?
When a conversation becomes a table, how many lines did not make it has to be visible too, because a missing row leaves no trace no matter how well the columns are split. If some lines did not match the format, the result shows how many under lines not read. That display is on screen only and does not go into the CSV.
Should I keep a KakaoTalk chat as a CSV or a PDF for a record?
A CSV is a format anyone can edit in a spreadsheet, which is good for sorting and analysis and exactly wrong where the point is to show the original unchanged. If you need the chat as a record, use the evidence markings of KakaoTalk Chat to PDF. The SHA-256 of the file you added is printed on the cover, so the same file can be checked against it later.
Can KakaoTalk Chat to CSV split a group chat by person?
Yes. A group export is read as it is, and the sender column carries each name as written, so you can filter it to one person in a spreadsheet or build a pivot table of messages per person. A real group export was checked against hand-counted row totals and the per-sender distribution.
What does KakaoTalk Chat to CSV do when it cannot read a single message?
It reads the head of the file to decide the format, and when no format fits it says the format was not recognised. A file whose format was recognised but from which no message could be read is treated as a failure too, so that you are never handed a CSV with a header row and no data. When it does read, the row count and the recognised format appear on screen, and you can check the first 50 rows in the preview before downloading.
Does KakaoTalk Chat to CSV read every PC time format?
Yes. Android, iPhone and PC write the export differently, and PC alone varies the time between a Korean 12-hour form, 15:24 and 3:24 PM. This tool tells those shapes apart on its own, including files exported with the English interface, and names the format it recognised on screen. A ZIP or an emailed .eml goes in as it is, and the chat file inside is found and read.