How to Fix a Blurry Picture
Soft edges narrowed on your own graphics card, in about a tenth of a second. Free, unlimited, no sign-up — and this page publishes how much it recovers and which blur it cannot touch.
- No sign-up
- No watermark
- Unlimited
- Free tools never upload
Drop images here
One at a time
Checking what this device can do…PNG, JPEG and WebP all open. Output is PNG, at the size you gave it.
How to fix a blurry picture, free and without an upload
Drop the photo into a browser tool that processes it locally. FreeUpscaler decodes it in the tab, runs a shock filter that narrows every soft edge on your own graphics card, and hands the picture back at its original size — no upload, no account and no watermark.
It recovers part of the definition, not all of it. Measured on fifteen photographs blurred by a known amount, the default setting brought back 14.3% of the edge definition the blur removed, and 20.3% at full strength.
Whether it helps at all depends on which kind of blur you have. The four kinds and their prognoses are the third block on this page, not a footnote.
What does fixing a blurry picture actually look like?
One photograph, softened by a known amount, run through this page's method, next to the frame it came from. All three panes are 100% crops straight out of the evaluation run.
- 14.3%
- Of the lost edge definition recovered
- mean over fifteen photographs
- 0.1 s
- For a four-megapixel photo
- no download, laptop GPU
- 0
- Files uploaded
- everything stays in the tab
- ∞
- Free runs
- no credits, no watermark
As it arrived · soft — edge gradient 23.2
Fixed · default setting — 25.5
The original · what the blur removed — 36.6
How much clearer, on what kind of photo?
Fifteen photographs, each softened by the same amount, each run through both available methods, each scored against the picture it came from. The unit is mean Sobel edge gradient; higher is closer to the original.
| Test photograph | Blurred input | Fixed | The original | Gap recovered◆ | Trained option |
|---|---|---|---|---|---|
| Family group, 1929 | 22.1 | 27.5 | 45.4 | 23% | 5.9% |
| Street, shop signage | 23.2 | 25.5 | 36.6 | 18% | 1.0% |
| Conference group, 2024 | 37.5 | 42.2 | 64.8 | 17% | 8.5% |
| Family group, 1900 | 27.4 | 31.1 | 49.7 | 16% | −0.7% |
| Street, bicycles | 22.4 | 25.8 | 44.1 | 16% | 5.7% |
| Portrait, studio | 20.5 | 23.2 | 37.4 | 16% | 3.8% |
| Wedding, outdoors | 28.2 | 33.8 | 65.5 | 15% | 7.1% |
| Street, houses | 26.9 | 31.2 | 56.7 | 15% | −2.3% |
| Portrait, 1940s archive | 10.7 | 11.8 | 18.6 | 14% | −2.5% |
| Food, pastries | 15.9 | 19.1 | 39.0 | 14% | 2.9% |
| Graduation, crowd | 13.6 | 14.4 | 19.6 | 13% | −2.7% |
| Food, cake | 17.1 | 19.4 | 35.9 | 13% | 3.7% |
| Portrait, official | 11.5 | 12.3 | 19.5 | 10% | −7.4% |
| Graduation, indoors | 9.8 | 10.3 | 15.4 | 9% | 1.0% |
| Food, shallow focus | 6.5 | 6.9 | 13.2 | 7% | −10.6% |
Every row is positive, and none of them is large. Against leaving the photograph alone, this method was ahead on all fifteen on both fidelity measures — there is no case in the run where pressing the button made things worse. But the spread runs from 7% to 23%, and the top of that range is still a picture that is visibly softer than the one the blur was applied to. Anyone promising to “unblur” a photograph is describing the right-hand column, not the middle one.
Where yours lands depends on what is in it. The pictures that recovered most are full of genuine edges at photograph scale — faces at a distance, printed lettering, spokes, fabric weave. The bottom row is a plated-fish close-up shot at a wide aperture: most of the frame is deliberately out of focus, so there are very few real edges for an edge-narrowing filter to act on. A photograph whose subject was meant to be soft cannot be made sharp, and 6.7% is what that looks like as a number.
The last column is the trained restoration network that also qualified for this page, and it is here so the choice not to offer it is checkable rather than asserted. Its median is 1.0% and six of its fifteen results are negative — it made an already-blurred photograph very slightly blurrier. On the structural measure it came out below the untouched input on eleven of the fifteen. It is trained on animation; a photograph’s soft edges are not the drawn lines it learned to rebuild.
Measured 31 August 2026 on an Apple M-series GPU in Chrome — fifteen photographs, two conditions, two methods, 60 runs, plus three sweeps of the strength setting. The full record is in the repository as corpus/findings/fix-blurry-photo.md.
First work out why it is blurry — it decides whether anything will help
Every tool in this category promises to unblur anything. That is not how images work. There are four common causes and they have very different prognoses; telling them apart takes about ten seconds and saves you a download.
- Improves a lot
Low resolution
Sharp but small — a messaging-app copy, a thumbnail, an old phone picture. All the captured detail is present, it just has too few pixels to show it. This is the one case where the right tool is not this page: enlarge it instead, because reconstructing at a larger size is what recovers apparent detail.
- Improves a lot
Genuine softness at the right size
A lens that is not especially sharp, focus a little off, a conservative camera profile, or a copy that has been through one compression too many. This is the case the page exists for, and the case the 14.3% figure was measured on. Pictures with real edges in them — faces, fabric, lettering — sit at the top of that range.
- Improves some
A photograph that was not really blurry
It will still change, and not for the better: on unblurred controls the default pushed edge contrast 14% past the original's own and moved measurably further from it in pixel terms. Turn the Deblur dial down to Light, which stays within 6%, or leave the picture alone.
- Improves some
Deliberate shallow depth of field
A portrait or food shot where the background is meant to be soft. There is nothing wrong with the photo and very little for the filter to act on — the lowest result in the whole run, 6.7%, was exactly this. Sharpening it mostly sharpens the grain in the blurred area.
- Cannot be fixed
Motion blur
The camera or the subject moved while the shutter was open, so every point of the scene was smeared along a line across many pixels. The test for it takes one look: if a stationary part of the frame is sharp and only the moving part is smeared, it is motion blur. Sharpening makes the smear itself crisper. That information is gone, not hidden.
- Cannot be fixed
Badly missed focus
The lens was focused somewhere else entirely. Each point became a disc rather than a point, and discs from neighbouring points overlap — which means the original values cannot be separated out again by any filter. Anything that hands back a convincing sharp version has invented the detail, and it is no longer your photograph.
If the before/after slider at the top of this page shows no real improvement on your file, one of the last two is almost certainly why — and no other tool will do better on it, whatever its landing page says. That is worth knowing before you upload a personal photograph somewhere to find out.
What it costs when the picture did not need it
Half the evaluation was a control group: the same fifteen photographs with no blur applied at all, run through the same settings. This is where a deblurring filter has nothing to gain and everything to lose, and it is the half a marketing page would not run.
A shock filter works by pushing the two sides of every edge apart — dark side darker, light side lighter — until the transition between them is narrow. On a soft edge that recovers definition. On an edge that was already correct it keeps going, past where the original was, and the result reads as brittle: outlines that look drawn on, grain that has been promoted to texture.
On photographs that were not blurred, the default pushed edge contrast 14.0% past the original’s own and dropped from 45.2 dB to 31.2 dB against it. Light stays within 5.6% and holds 34.7 dB. That is the whole argument for looking at your picture before you press anything.
It is also why the default is Medium rather than Strong. Strong recovers six points more on genuinely soft pictures, and overshoots by 20.5% on ones that were fine. A default has to be right for the person who is not sure, and that person’s photograph is more often nearly-fine than badly soft.
Not blurred, untouched · the control — 34.8
Same picture, default applied · 42.1 — past the original's own 36.6
| Deblur setting | Recovered on a soft photo◆ | Distance from the original | Overshoot when nothing was wrong |
|---|---|---|---|
| Light | 8.2% | 34.71 dB | +5.6% |
| Medium — the default | 14.3% | 31.21 dB | +14.0% |
| Strong | 20.3% | 29.53 dB | +20.5% |
There are three steps and no fourth. The dial buys iterations of the edge-narrowing pass and the engine caps them at three, so Strong is genuinely the ceiling of this method rather than the top of an arbitrary scale. Every one of the fifteen photographs improved monotonically across the three, with no photograph peaking early — which is a polite way of saying the ceiling is the method, not the tuning.
Why do iPhone photos come out blurry, and can this fix them?
Usually it can not, and the reason is worth more to you than the tool is: most blurry phone pictures are motion-blurred, which is the one case nothing recovers. All four causes below are fixable at the source, and three of them only at the source.
- A smudged lens. The single most common cause, and the least suspected. Pocket lint and fingerprints scatter light across the whole frame, which softens everything evenly. Wipe the lens and take the shot again — the difference is larger than anything on this page can produce.
- Low light. In dim conditions the camera holds the shutter open longer, so ordinary hand movement smears the picture. This is motion blur, and it is the case marked “no” above. Brace the phone against something solid, or add light.
- Focus on the wrong subject. The phone chose a face in the background, or the railing in front. Tap the subject on screen before shooting to lock focus there. If the shot is already taken, this is badly missed focus and it does not come back.
- It arrived through a messaging app. WhatsApp, Messenger and similar re-compress and shrink what they carry, so the copy on your phone is a degraded version of a picture that is probably still fine on the sender’s. Ask for the original first. If you cannot get it, this is the case that does improve here — and if the copy is also much smaller than the original, enlarging it will do more than deblurring it.
The short version: if the photograph is soft, drop it in above. If it was taken handheld in a dark room, retaking it beats every piece of software there is, including this one.
What the method actually does to your picture
One pass, no network, no model download — and no black box, because the whole thing is a few lines of shader maths.
A shock filter, run up to three times
At every pixel it asks which side of an edge that pixel is on, by looking at the curvature of the brightness around it, and then moves it a small step towards whichever side it already belonged to. Repeat that and a gradual ramp between dark and light becomes a short one. The step is deliberately small — a single large step would quantise the edge into a staircase, while small repeated steps converge on the same place smoothly.
What it cannot do follows directly from that description: it moves values that are already there. It has no way to separate two overlapping discs of light back into two points, which is what recovering real defocus would require, and it has no notion of what your photograph is of, so it cannot invent what should have been there.
Why we do not call it AI
Because it is not one, and the distinction is useful to you rather than pedantic. There is no model to download, so it starts the instant you drop a file in. It works offline. And it will never hand back detail that was not derived from your own pixels — which is exactly the failure mode of the generative tools in this category, where a face comes back sharp and subtly not the same face.
The full pipeline, and where the trained networks are used on other pages, is on how it works.
Three steps, no account
Everything happens on this page. There is no upload to wait for and no email to confirm.
- Step 1
Drop your photos in
Drag one picture or a dozen onto the box above, or click to browse. The page checks what your device can handle before you choose a file, so you learn about limits up front rather than half-way through.
- Step 2
Judge it on real detail
Drag the before/after slider over eyelashes, fabric, lettering or foliage — not over a flat wall, where by definition nothing changes. If the after side looks brittle rather than defined, turn the Deblur dial down to Light and let it re-run.
- Step 3
Download it
No watermark, no sign-up, no cap. The output is a PNG at the same dimensions you supplied, so it drops straight back into wherever the original came from.
Free without an asterisk
Genuinely unlimited
No credit counter, no daily cap, no queue. Your GPU does the work, so the hundredth photo costs us exactly what the first one did: nothing.
Nothing is uploaded
Your photograph is decoded and processed inside this tab. There is no server in the loop and nothing for us to store — which matters more here than on most pages, because the pictures people want to rescue are personal ones.
No account, no watermark
No email, no sign-in wall before the download button, and no logo stamped across your result.
Works offline
Load the page, disconnect from the internet, and it still runs. That is the proof the privacy claim is real, and it takes ten seconds to check.
Starts instantly
Shader maths, not a network — nothing to download, no queue, and a four-megapixel photo finishes in about a tenth of a second.
The limits are published
Both what it recovers and what it costs when it should not have been used are on this page as measured numbers. That is the half of a measurement a marketing page leaves out.
Not quite the job you have?
Upscale a video
For footage that looks soft because it is small. Reconstructs at 2×, 3× or 4×.
OpenEnhance photo quality
For a picture that is soft because it is small. Enlarging is what recovers apparent detail; this page does not change your dimensions.
OpenFix a blurry video
The same filter frame by frame, where it measured more than twice the recovery it gets on stills.
OpenFix a pixelated photo
For blocky squares and stair-stepped edges rather than soft ones — a different damage with a different answer.
OpenDenoise an image
Edge-preserving noise reduction, with the trade-off published as measured numbers.
OpenChange image DPI
Resamples the pixels first, then writes a real DPI — unlike metadata-only converters.
OpenFrequently asked questions
Can a blurry picture actually be fixed?+
Partly, and how much depends entirely on why it is blurry. On fifteen photographs softened by a known amount, this page recovered an average of 14.3% of the edge definition the blur removed, and up to 23% on the best of them — a real improvement that is a long way from restoring the original. Motion blur, where the camera or subject moved during the exposure, and badly missed focus cannot be recovered at all: the light from each point of the scene was physically spread across many pixels at capture time, and no filter puts it back.
How do I fix a blurry picture for free?+
Drop the photo onto this page. It is decoded in your browser, sharpened on your own graphics card by a shock filter that narrows soft edges, and handed back — no upload, no account, no watermark and no limit on how many you do. A four-megapixel photo takes about a tenth of a second, and nothing has to be downloaded first.
How much clearer does the picture actually get?+
Between 7% and 23% of the lost edge definition came back across the fifteen test photographs, averaging 14.3% at the default setting and 20.3% at Strong. Where yours lands depends on what is in it: pictures with plenty of real edges — faces, fabric, lettering, foliage — recover most. A photograph whose subject is a shallow-depth-of-field close-up recovered the least at 6.7%, because there were few edges to work on in the first place.
Why is there only one method to choose from?+
Because only one of the two that could run here does anything measurable. The alternative is a trained restoration network, and across the fifteen test photographs it recovered 0.9% of the lost definition — a median of 1.0%, with six of the fifteen going slightly backwards. It is trained on animation, and the soft edges in a photograph are not the drawn lines it learned. Offering a 924 KB download that measures as nothing would be a worse picker, not a richer one. It is still offered on the blurry-video page, where the same network measured 13.4%.
Will it ruin a photo that was not really blurry?+
It will overdo it, which is why the number is published. Run on photographs that had not been blurred at all, the default setting pushed edge contrast 14% past the original's own and moved the pixels measurably further from it. If your picture is only slightly soft, turn the Deblur dial down to Light, which stays within 6%. The honest first step is deciding whether the picture is actually blurry before you press anything.
Is there a strength setting?+
One, in the options popover, labelled Deblur, with three working steps. Each notch runs the edge-narrowing pass one more time, and each does something measurable: 8.2%, 14.3% and 20.3% of the lost definition recovered at Light, Medium and Strong. The cost rises with it — on a photo that needed nothing, the same three settings overshot the original by 6%, 14% and 21%. If the result still looks soft, go up; if it looks brittle, go down.
My iPhone photos come out blurry. Does this fix them?+
It helps if the photo is soft, and not at all if it is motion-blurred — and on a phone the second is more common than the first. Most blurry iPhone pictures have a cause that is fixable at the source: a smudged lens, low light forcing a long exposure, or focus locked onto the wrong thing. If the photo arrived through a messaging app it was re-compressed and shrunk, which is a softness this page does improve. There is a section on this below.
Is it free, and is there a watermark?+
Free with no limit and no watermark. No account, no credit counter, no export cap. Your own GPU does the processing, so an extra photo costs us nothing — hosted tools meter you because they rent hardware by the minute and have a bill to recover.
Do you upload my photo?+
No. The image is decoded and processed inside your browser tab. You can verify it in ten seconds: load this page, disconnect from the internet, and fix a photo anyway. It still works, because there was never a server in the loop.
Which format should I use?+
PNG, JPEG and WebP all open. Output is always PNG, because re-encoding a freshly sharpened image as JPEG would immediately give back some of the definition you just gained.
Should I fix the blur or enlarge the picture?+
If the picture looks soft because it is small — a thumbnail, a messaging-app copy — enlarge it instead: reconstructing at a larger size is what recovers apparent detail, and sharpening happens as part of that pass. This page is for a photograph that is already the right size and still looks soft. It does not change your image's dimensions at all.
More free tools on this site
Fix a blurry picture right now
No account, no watermark, no limit. Drop a photo in and it is processed on your own machine.
Choose a file