Publishing Shorts Everywhere:
A Platform-by-Platform Guide
You made a 45-second news Short. The caption you wrote for YouTube is too long for TikTok, the one for Instagram reads like a press release on LinkedIn, and the one for X does not fit in X. Rewriting the caption eleven times by hand is how creators burn out on the final, lowest-leverage step of a video they already made. The video is done. The only thing left is to say the same thing, well, in eleven different ways.
This guide is about how to do that. The conventions each platform rewards, the platforms that will accept a direct post and the ones that will not, and how to schedule across them without a spreadsheet. It is the pair to the hooks guide — that one is about the first two seconds of the video, this one is about the sentence that gets a viewer to open it.
A caption is a piece of writing in the voice of a specific platform. The same story told in the length, the hashtag count, and the tone that TikTok rewards is a different sentence than the one LinkedIn rewards. Writing it once and copy-pasting it eleven times is how the same video reaches nobody.
Why each platform needs its own caption
The obvious difference is character count — TikTok caps at 150, X at 270, Instagram at 2,200. That is not the interesting part.
On TikTok, the caption is a second hook. It sits on top of the video, in tiny text, and most viewers read it before the video starts. A TikTok caption that opens with a summary loses; one that opens with a curiosity line holds. On Instagram, the caption is a companion, usually read after the video, and the first line is the only one that shows before the “more” cut. On X, the caption is the entire post — the video plays inline, but the sentence is what gets read, quoted, and replied to. On LinkedIn, the caption is a professional framing device: the same news, with a lens that respects that the reader is at work.
Hashtags are the same story in miniature. Instagram rewards 5–10 relevant hashtags; TikTok rewards 3–5 and punishes a wall of them; LinkedIn uses 3 and reads as spam at 10; X uses 1–2 and looks amateur at 4; Threads and Bluesky barely use hashtags at all. A caption that follows the X convention on Instagram looks like a mistake, and a caption that follows the Instagram convention on X gets cut off at the “more”.
The good news is that the differences are learnable. The bad news is that they are not the differences a person reaches for by instinct. The person who wrote the YouTube caption was optimising for the YouTube caption; the same person writing the Instagram caption from scratch would write a different sentence, because they would be reading a different platform’s feed while they wrote it. This is what a tool is for: the model writes each platform’s caption in that platform’s voice, from the same story facts, in the same call.
The eleven platforms and what each wants
The table below is the current set of platforms Autonom writes captions for. The character limits, hashtag counts, and tones are the ones the tool enforces, so a caption written for TikTok will not arrive on Instagram cut off, and one written for LinkedIn will not arrive on X with 800 characters of body missing. The worked examples are the same story — a single fact from a single article — written in each platform’s voice.
| Platform | Limit | Hashtags | Tone | Worked example |
|---|---|---|---|---|
| YouTube Shorts | 5,000 | 3 | Description that adds context and sources; the title carries the hook. | “The EU has set tariffs of up to 45% on Chinese electric vehicles. Here’s what Brussels says the duties are for, and why Beijing says it will challenge them.” |
| Instagram Reels | 2,200 | 6 | A first line that stops the scroll, 2–3 short lines, a question, then the hashtags. | “The EU just put tariffs on Chinese EVs — up to 45%.\n\nBrussels says state subsidies gave Chinese makers an unfair edge. Beijing says it will challenge the decision.\n\nDo tariffs actually work, or do they just get passed to buyers?\n\n#EU #China #EVs #TradePolicy #Tariffs #Autonom” |
| Facebook Reels | 600 | 3 | 2–3 plain sentences and a question. Few hashtags. | “The EU has put tariffs of up to 45% on Chinese electric cars. Brussels says subsidies gave Chinese makers an unfair advantage; Beijing says it will challenge the decision. What do you think this does to EV prices here?\n\n#EU #China #ElectricVehicles” |
| TikTok | 150 | 4 | One curiosity-led line, under 150 characters including the hashtags. | “Up to 45% on Chinese EVs — and Beijing is not happy #EU #China #EVs #Tariffs” |
| X | 270 | 2 | One sharp take under 270 characters. No link. At most two hashtags. | “The EU just set tariffs of up to 45% on Chinese EVs. Brussels: unfair subsidies. Beijing: we’ll challenge it. The interesting part is who pays if neither side blinks. #EU #China” |
| 1,100 | 3 | Professional, 3–5 short lines leading with the analytical point, ending on a question. | “The EU has formally imposed tariffs of up to 45% on Chinese electric vehicles.\n\nThe stated rationale is state subsidies; the real question is what happens to European EV prices if Beijing retaliates. Two things to watch: whether member states stay united, and whether Brussels opens a negotiation channel in parallel.\n\nWhat would a managed de-escalation look like here?\n\n#Trade #EU #China” | |
| Threads | 480 | 1 | Conversational, like a thought you had out loud. | “Up to 45% on Chinese EVs. Brussels says subsidies, Beijing says it will challenge. Genuinely don’t know which side blinks first here.” |
| Bluesky | 290 | 1 | One clear sentence or two, under 290 characters. | “EU tariffs of up to 45% on Chinese EVs are in. Beijing says it will challenge. Everyone waiting to see who moves first on retaliation.” |
| Telegram | 900 | 3 | 3–4 sentences, plain text, a question at the end. | “The EU has imposed tariffs of up to 45% on Chinese electric vehicles. Brussels says state subsidies gave Chinese makers an unfair edge. Beijing says it will challenge the decision at the WTO. What this does to European EV prices is the question nobody is answering yet.\n\n#EU #China #Trade” |
| Rumble | 1,000 | 3 | A short description: what happened and the angle. | “The EU has imposed tariffs of up to 45% on Chinese electric vehicles. This short looks at what Brussels says the duties are for, what Beijing says it will do about them, and why both sides are hedging. #EU #China #Tariffs” |
| Mastodon | 480 | 2 | Plain, conversational, under 480 characters, hashtags at the end. | “The EU has imposed tariffs of up to 45% on Chinese electric vehicles. Brussels cites unfair subsidies; Beijing says it will challenge at the WTO. Worth watching whether member states stay united if retaliation comes. #EU #Trade” |
The limits and hashtag counts are not the whole story, but they are the mechanical part of the convention, and getting the mechanical part wrong is the fastest way to look like a tool sprayed a caption across every platform. The rest is the tone, which is why each row in the table above has a worked example rather than just a character count — the shape of the sentence matters as much as how long it is. If you want to see how the same story becomes the video itself, the hooks guide covers the opening line, and the voice guide covers how the narration is timed to fit.
What can post directly and what can’t
There are two ways a caption leaves the tool: post it directly, or copy it and use the platform’s own uploader. Which one you get depends on whether the platform lets a browser post to it, and that depends on the platform, not on the tool.
Two platforms can post directly from the browser: Telegram and Mastodon. Both have open APIs that accept a video upload and a caption from a user’s own token, so if you have set up a bot token for Telegram and an access token for Mastodon, the tool can send the finished video and its caption straight to the channel or account. No download, no copy-paste, no separate uploader.
The other nine need a copy-paste, or a share from your phone. YouTube, Instagram, Facebook, TikTok, X, LinkedIn, Threads, Bluesky, and Rumble do not accept video uploads from a browser session the way Telegram and Mastodon do. Some of them (X, Bluesky, LinkedIn) will open a post composer with the caption already in it via a share URL, which gets you most of the way there — the caption is filled in, you upload the video, you press post. Others (YouTube, Instagram, TikTok, Facebook, Rumble) require the platform’s own uploader: download the file, open the app or the studio, upload the video, paste the caption. That is more steps, and it is not a defect of the tool; it is what those platforms allow.
The browser release also supports the phone share sheet. If you are on a phone, you can share the finished MP4 straight from the browser to the Instagram app, the TikTok app, or the YouTube app, with the caption pre-filled by the platform’s own share intent. It is the least-friction path that does not require a server to do the posting on your behalf.
In every case, the video itself is the same file. The voice guide covers how the audio is produced; the hooks guide covers the opening line of the video and of the caption. Publishing is the step where the two meet: the sentence that opens the video and the sentence that opens the caption have the same job, and they should be recognisably the same idea in two different voices.
Scheduling without pain
Scheduling Shorts is the part of publishing that a spreadsheet makes worse. The instinct is to post everywhere at the same time, because posting everywhere at the same time is what a single action can do. That instinct costs you reach, because the platforms do not agree on when their audiences are watching, and posting to five platforms at once means posting to each of them at the wrong time for four of them.
Each platform has its own native scheduling, and it is better than a third-party scheduler for that platform’s own algorithm:
- YouTube Studio lets you schedule a Short for a specific date and time, and it is what the algorithm sees as the publish event. If your channel’s viewers are most active on weekday mornings, schedule for weekday mornings, not “when I finish making the video”.
- Meta Business Suite covers Facebook and Instagram in one place, with per-platform times. This is the closest thing to a cross-platform scheduler that does not require a paid third-party tool.
- TikTok Studio has native scheduling up to 10 days ahead, which is enough for a weekly cadence and not enough to build a month of queued content. Plan to re-fill weekly.
- LinkedIn has native scheduling inside the post composer, which most people never find. It works; it just isn’t where you would look for it.
- X has native scheduling in the web composer, but only on the web (not in the app). It also supports threads with a scheduled first post, which is the only sensible way to schedule a thread.
The better version of this problem is to space the platforms out, rather than posting everywhere at once. Publish to YouTube first, where the algorithm rewards a clean first 24 hours. Publish to Instagram and TikTok a few hours later, in their own peak windows. Publish to X, LinkedIn, and Threads the next morning, in their own local-time peaks. The story is already made; the only thing changing is when a viewer sees it, and a viewer who sees it in their own feed at their own peak is a viewer who might actually watch.
The simplest workflow that works for a solo creator: publish everything by hand once a day, at a time you can sit down and post to each platform in turn. It takes fifteen minutes, it uses each platform’s own scheduler (or no scheduler at all), and it avoids the two failure modes of a fully automated cross-post — that the caption is generic because it was written once, and that the timing is wrong because it was chosen once.
Here is a generic caption, of the kind that gets copy-pasted to every platform, next to the caption the same story should have on Instagram. Both contain the same fact. Only one was written for the platform it is on.
— Instagram-specific: a first line that stops the scroll, two short lines of body, a question that invites a reply, six hashtags at the end.
The second caption is longer than the first. That is not a failure; Instagram wants a caption that a viewer reads after the video, and a viewer who has just watched a 45-second story is willing to read four lines about it. The first caption is the shortest version of the same story, and it is short enough to be useless on every platform except X — where it is also too generic. The lesson is not “longer is better”. The lesson is “fit the platform”, and the platform is the whole reason a caption exists.
One story, eleven captions, written in each platform’s voice.
Autonom Newsreel writes the caption for every selected platform from the same story facts, in that platform’s length, hashtag convention, and tone. Post to Telegram and Mastodon directly; download the MP4 and copy the caption for the rest. Free to run on your own API keys.
Open Autonom Newsreel →Related guides
Frequently asked questions
How do I post a YouTube Short to Instagram?
Two ways. The easy way, if you are on a phone: open the finished MP4 in your browser or camera roll, tap Share, and choose Instagram. The Instagram app will open with the video already attached. The longer way: download the MP4, open Instagram, tap Create, upload the video as a Reel, and paste the Instagram-specific caption. The caption matters more than the path: a caption written for YouTube, pasted into Instagram, reads like a YouTube caption pasted into Instagram. Write it for Instagram.
Why can’t I post directly to YouTube, Instagram, and TikTok from the browser?
Because none of those platforms allow a browser session to upload a video to them without an approved developer app behind it. Telegram and Mastodon have open APIs that accept a video upload from a user’s own token; the others do not, or require an approved application and an OAuth flow that is not part of a free browser tool. Copy-paste and the phone share sheet are the honest paths, and they work everywhere.
What is the best time to publish a Short?
When your own audience is watching, which is a per-channel question and not a universal answer. The industry default — weekday mornings, local time — is a starting point, not a result. Check your YouTube Studio analytics for “when your viewers are on YouTube” and use that window. On Instagram, TikTok, and LinkedIn, the native analytics tell you the same thing per platform. The mistake is picking one time and using it for all of them; the platforms do not share an audience, so they do not share a peak.
Should I use a third-party scheduler like Buffer or Later?
For a solo creator publishing a handful of Shorts a week, the native schedulers are usually better — they post through the platform’s own API, which the algorithm treats as a normal publish event, and they don’t have a paid tier between you and your own content. For an agency publishing dozens of videos a week across many accounts, a third-party scheduler saves enough time to justify the cost. The decision is volume, not preference.
Does the same video need a different caption on every platform?
The same video needs the same story told in the platform’s voice, which means a different sentence for each one. It does not mean a different angle — the angle is the video’s job. The caption’s job is to translate the video’s opening line into the platform’s length, hashtag convention, and tone, so a viewer who sees it in their feed recognises the same story they would see if they had found the video on any other platform. Autonom writes all eleven from the same story facts, in the same call, so a viewer who follows you on three platforms gets the same story in three different voices, not three different stories.
What happens if I don’t have time to write eleven captions?
Write three and copy-paste the rest with the length and hashtags adjusted for each. Three is the minimum that covers the platforms with the biggest convention gaps: one for video-first platforms (YouTube, Rumble), one for scroll-first platforms (Instagram, TikTok, Threads, Bluesky), and one for text-first platforms (X, LinkedIn, Mastodon, Telegram). The captions in the same row of that table are the ones that can honestly be variants of each other; the ones across rows cannot. That is why the table has five columns and not one.