Inspiration
Every post you have saved, any network, any format. Studio, then Inspiration.
Postlyra does not go looking for posts on your behalf. It works on the ones you save: it reads them, breaks down why they worked, and turns them into ideas, hooks and scripts with the source attached.
Saving a post
There are three ways in. Each row records which one it came from.
| Way in | Shown as | Best for |
|---|---|---|
| The Chrome extension | Extension | Finding outliers in a feed you are already scrolling, then saving them |
| Save by link in Studio | Manual | A post someone sent you, or one you have open in another tab |
| The phone's share sheet | Phone | A post you meet in an app |
Saving the same post twice never makes a duplicate. The second save fills in the same row.
Finding outliers with the Chrome extension
The extension ranks the feed you are already looking at. It works on Instagram, Facebook, Threads, X, TikTok, YouTube and LinkedIn. In Chrome it is named Viral Post Finder, Content Ideas & Research by Postlyra.
- It never reorders the page. The platform keeps its own layout and its own scroll. The ranking happens in a side panel, and each recognised post gets a small badge in place.
- Badges show the post's counts and, above 3x, its outlier multiple. Above 10x the badge is marked hot. A badge only appears on posts still on screen: Instagram and Threads drop posts from the page as you scroll past, so older ones are in the side panel list but no longer badged.
- The side panel ranks by views, likes, comments, shares, engagement, outlier multiple or date, in either direction, and filters by author or caption. This feed lists what was collected on the surface you are on (one profile, one feed). All feeds lists every surface it has kept, scored the same way.
- Ads are hidden by default, from the list, the page badges and the CSV. A setting shows them again, labelled as ads, which is what you want when a competitor's ad is the thing you are studying. Ads never count toward the baseline either way. Ad detection is per platform and best effort: LinkedIn and Threads expose no marker the extension reads yet, so their ads can still appear.
- Export as CSV is in the panel's ... menu, with Save all to Postlyra and the two clear actions.
- A Postlyra account is optional. Ranking, badges and CSV work without one. You only sign in to save posts into Studio.
What "outlier multiple" means here
The multiple compares a post with the median of the same creator's posts the extension has collected on that page. It is a measure of how far a post beat its own creator's usual result, not of how big it is. A post with 50,000 views from a creator who normally gets 5,000 is a 10x. The same 50,000 from a creator who normally gets 500,000 is nothing.
The number it compares is views on YouTube and TikTok, and likes plus comments on Instagram, Threads, X and LinkedIn. X shows views, but the extension ranks X on engagement anyway, because an impression count is what the algorithm chose to distribute, not what the post earned. Facebook's reels viewer does not expose a play count, so Facebook ranks on engagement too.
It is a median over what the extension has actually seen, so scroll a creator's profile for a while before you trust a multiple. A creator with three posts collected has a weak baseline.
Two limits to know about. Facebook sends a post's caption, counts and video in separate pieces, so counts can fill in a moment after a post first appears. And an Instagram profile grid publishes no per-post numbers at all, so those posts are collected but cannot be ranked. Instagram's feed and Reels tab are fine.
What it reads
It reads what your own browser already loaded. It watches the data the platform's own page fetches, plus the page itself, and adds no requests of its own to the platform. It never logs in to a platform for you. Collected posts are kept in the extension's own storage in your browser, per profile and per feed, and stay there until you clear them or save them.
Saving is the only step that sends anything to Postlyra. The extension sends the post's link, text, counts, thumbnail link and multiple, and for a video the link to its video file (and, on Facebook, its caption file). When you save a video, Postlyra downloads that video's audio or caption file and its thumbnail from those links, once, to transcribe it and keep the picture. Nothing else is fetched from the platform.
Saving from it
Save from a badge or from the list. Each post lands here as a saved post of its own format: video, carousel, image or text. In the extension's settings you can choose to analyse every saved post right away, on your AI credits.
Save by link
Save by link, top right of the page, opens a box: paste a link to a post, reel or video on YouTube, Instagram, Facebook, TikTok, Threads, LinkedIn, Bluesky or X, and it lands here. Paste a whole forwarded message and it still works, since the box picks the link out of it. A short link (vm.tiktok.com, lnkd.in, fb.watch, facebook.com/share/...) is resolved to the real one first.
Analyse after saving is a checkbox on the same box, off by default: check it and the AI breakdown runs right away, on your AI credits.
What actually lands depends on the platform, and this is the one path where Postlyra's servers do read the platform. A YouTube link comes back complete through YouTube's own API, transcript included. Every other platform gets whatever the public page shows a visitor who is not logged in: title, caption, picture, and the video file when the page exposes it, which starts the audio transcription described below. On Facebook and Instagram that is often nothing at all, because both put most of a post behind a login wall. Postlyra says so right after you save: the post is kept with just its link, and saving the same link again from the extension or the phone's share sheet fills in the full post on that same row.
Facebook videos and reels are the clearest case. A plain link often saves nothing but the link, so use the extension or the phone for those.
A post saved by link has no outlier multiple, because there is nothing to compare it with.
Sorting and filtering
Sort by Recently saved (the default) or Outliers. Filter by platform, format (video, carousel, image, text), minimum outlier score, published date range, whether it has been analysed, collection, and free text. The free text searches the title, caption, transcript and analysis.
Only posts saved with a multiple, which means from the extension, have an outlier score to sort or filter on. The rest sort last.
What you can do with one
| Action | Result |
|---|---|
| Fetch transcript / Transcribe audio | Pulls captions when there are any, otherwise transcribes the audio. Video only |
| Analyse | Breaks down why it worked: the hook, the storytelling, the format, the style and the takeaways |
| Turn into an idea | Adds it to your idea Library as is, no AI involved |
| Generate ideas | AI-mined content ideas from the analysis, into your idea Library |
| Generate hooks | AI-written hook variations, into Hooks |
| Save formula | Saves the hook's underlying structure as a reusable template, once it has been analysed |
| New script | Starts a script with this as reference. Video only |
| Draft a post | Writes a draft in your voice |
| Add to a collection | Files it in a collection |
Analysis runs on the transcript when a video has one, and on the written caption otherwise. A caption-only analysis says so.
Transcript status is none, fetched, or unavailable. There are two sources, tried in order. First, the platform's own caption track, free, no AI credits: YouTube always has one, and Facebook auto-generates one for many reels, which the extension picks up and sends along with the save. Only when there is no caption track does it fall back to speech recognition on the saved video's audio: Instagram and TikTok never publish one, so they always take this path, and so does a Facebook reel that never got auto-captioned. The audio path is an AI call charged against your AI credits, about $0.003 a minute, and it runs on its own right after the post is saved, while the video file link the extension or the phone sent along with it still works, so most of the time there is nothing to click. The button says Fetch transcript when a free source exists and Transcribe audio only when the audio is the only one there is, and it shows disabled with a note to save the post again when neither a caption track nor a media link ever arrived. A video longer than 10 minutes or bigger than 50 MB is reported unavailable and never billed, and a reel that is mostly music and on-screen text can come back with little or nothing worth reading, in which case the analysis falls back to the post's written caption instead.
On the phone
The mobile app has the same feed, under Inspiration on its home screen. Search it, filter by platform, format, source and whether a post has been analysed, and sort it. The phone's filters stop there: channel, minimum outlier score, date range and collection are Studio only.
Sharing a post to Postlyra from another app saves it here. The confirmation opens the row it just saved, rather than telling you to go and find it in Studio later.
Open a saved post on the phone and you get its caption or transcript, the full analysis, and three of the actions above: Fetch transcript / Transcribe audio, Analyse, and Turn into an idea. The rest of that table is Studio only. Analysing on a phone is the same AI call on the same credits and takes the same few seconds, so expect to wait on the button.
A working loop
- Scroll the feeds you care about with the extension open, and rank by outlier multiple.
- Save the few that made you stop.
- In Studio, open each one, fetch the transcript if it does not have one, then Analyse.
- If the insight transfers to your audience, Generate ideas.
- Triage those in Planning like any other idea.
Ten minutes, a couple of times a week. When you have analysed a handful, Reports looks for what they have in common.
Do not copy
Analysis tells you why something worked: the structure, the hook, the promise. That transfers. The topic usually does not.
Ideas generated from a post are grounded in your persona and pillars, not in the source's subject, which is the point.