Threads downloader, explained

JPEG, every time. This is the one part of the Threads photo downloader that surprises people, so it is worth explaining properly.

Threads stores uploaded photos as HEIC, and the URLs in a post's metadata still end in .heic. But those URLs also carry a conversion instruction, and Meta's CDN converts on the way out — what actually arrives over the wire is a JPEG. So the file is named .jpg, because that is what the bytes are.

As a backstop, the download proxy also checks the Content-Type of the response and names the file from that. If Meta ever changes how the conversion is signalled, you still get an extension that matches the contents rather than an unopenable .heic.

Copy link to this answer

The first one. Options are ordered largest first, so the top row is the biggest rendition Threads publishes for that item — that row is your HD thread download.

The smaller rows exist for when you want a lighter file — sending a photo in a chat, say, rather than archiving it. Sizes below 1080px on the long edge are not listed at all, because at that point they are thumbnails rather than downloads.

One caveat worth being straight about: the largest rendition Threads publishes is Threads' own copy, not the untouched file the uploader selected from their camera roll. No downloader can reach the latter, and any that claims to is guessing.

Copy link to this answer

Because Threads generates a square version of every photo for grid previews, and it sits in the same list as the full-frame sizes.

Choosing it means losing the top and bottom of a portrait photo, so those rows carry a small crop icon and say so when you hover. Rows without that icon are the whole frame.

Copy link to this answer

Not by us. We read the rendition list from the post's own public metadata and stream the one you pick straight through, byte for byte. No watermark is added, no frame is re-compressed, and no branding is burned in.

Threads usually publishes the same video at a few quality tiers. Duplicates are collapsed so you see the tiers you can actually choose between, largest first. Picking the top row gives you the original bitrate and resolution as Threads published it — an HD thread download with nothing altered in between.

Copy link to this answer

Descriptively, so a downloads folder stays navigable. The name is built from the author's handle, the post id, the position in a carousel, and the pixel size:

• just-adulting-DbP_Cu-gT7r-1440x1800.jpg
• just-adulting-DbP_Cu-gT7r-2-1080x1350.jpg

The size is in the name because a post now offers several of them, and three files with identical names would defeat the point of offering a choice.

Copy link to this answer

Two thresholds, and normal posts clear neither.

• Above roughly 150 MB the in-page progress bar steps aside and your browser's own download manager takes over — holding a file that size in a tab's memory is how downloads fail on phones.
• Above 512 MB the server declines to proxy the file at all. Threads does not publish videos anywhere near that, so this is a guardrail rather than a limit you will meet.

Copy link to this answer

No. Every Threads video download is piped from Meta's CDN through to your browser as it arrives. Nothing is buffered to disk, nothing is written to a database, and there is no archive of past downloads to leak or subpoena.

That is a structural property, not a promise about our intentions: the proxy has no storage layer to write to.

Copy link to this answer

• Resolved post metadata — caption, handle, media URLs, dimensions — in memory for about ten minutes, so looking up the same link twice does not hit Threads twice. It is gone on expiry or restart.
• Rate-limit counters, in memory, cleared at the end of each one-minute window.
• Ordinary web server access logs, kept for as long as the host's default, for abuse investigation.

No media, no accounts, no download history. The privacy policy spells out the same list in more detail.

Copy link to this answer

Not for the download itself. Your browser asks our server, our server asks Meta, and the bytes come back through us — so Meta sees a request from this server, with no referrer or fingerprint of yours attached.

One deliberate exception: the preview thumbnails in the results grid are loaded by your browser directly from Meta's CDN, with a no-referrer policy. It keeps the page fast and the bandwidth honest. Blocking images for this site removes that connection without affecting downloads.

Copy link to this answer

No analytics scripts, no advertising trackers, no third-party pixels, and no cookies at all.

One value is written to your browser's local storage: whether you prefer the light or dark theme. It never leaves your device.

Copy link to this answer

The error message on screen names the specific cause. In order of how often they come up:

• The link points at a profile or a feed rather than at one post.
• The post is private, follower-only, or age-restricted, so its media is not publicly served.
• The post was deleted, or the id in the link has a typo.
• The post has no downloadable media — text-only posts, quote posts, and link previews have nothing to save.
• Threads served this server a challenge page instead of the post, which is usually temporary.

If the message says something went wrong on our side rather than naming one of the above, waiting a minute and retrying is genuinely the right move.

Copy link to this answer

Meta's CDN URLs are signed and time-limited. If you leave a results page open for a while, the links behind those buttons go stale even though the page still looks fine.

Run the lookup again to mint fresh ones. Nothing is wrong with the post or with your link.

Copy link to this answer

There is a per-address cap of twenty lookups a minute. It exists to stop one scripted client from burning through Meta's tolerance and getting this server throttled for everybody.

The message tells you how many seconds remain. Downloading files you have already resolved does not count against the cap — only new lookups do.

Copy link to this answer

No, and this is not a limitation we intend to work around. Those posts are not publicly served, so reaching them would mean impersonating a logged-in follower.

If it is your own private post, the Threads app can save your media directly.

Copy link to this answer

The upstream request is given twelve seconds before being abandoned, so a hung connection cannot leave you staring at a spinner.

This is almost always transient. Try again; if it persists for several minutes, Threads is likely having a bad time rather than your link being wrong.

Copy link to this answer

Not without their permission. The person who uploaded it owns it, and a download button does not transfer that.

What this tool is for: saving your own posts, and keeping a personal offline copy of things you are allowed to keep. Re-uploading someone else's video as your own is a copyright problem regardless of how you obtained the file.

Copy link to this answer

Only with a licence from the rights holder. Advertising, client work, paid content, and monetised compilations all need permission, and "it was public" is not a defence.

If you are unsure whether your use qualifies, that is a question for a lawyer rather than for a download page.

Copy link to this answer

There is nothing here to take down — no copy of your media exists on this server. Files stream from Meta's own CDN to whoever requested them and are never retained.

If content of yours is being hosted or redistributed somewhere else, that host is the effective place to file. Our copyright policy explains what we can and cannot act on.

Copy link to this answer