Request access

Instagram API 2026: saves, shares and reposts

Reposts had never been counted anywhere before April 2026. That is the real story in an update that most coverage summarised as Meta opening up saves data, which is not what happened.

In short

Did the April 2026 API change give me anything I did not have?

On 22 April 2026 Meta added reposts_count, saved_count and shares_count as fields on the Instagram media endpoint for apps using Facebook Login, across all API versions. Saves and shares were already obtainable through the insights edge, so for those two the change is about how they are fetched rather than whether they can be. Reposts are genuinely new and had no prior equivalent. None of the three became available for accounts you do not manage. The practical effect is more reliable collection on mixed feeds and one new distribution signal worth adding to reporting.

An archive cabinet with two drawers already open and a third just unlocked by an amber key: reposts newly counted

On 22 April 2026 Meta added three fields to the Instagram media endpoint for apps using Facebook Login:

  • reposts_count - instances the media was reposted
  • saved_count - times the media was saved
  • shares_count - times the media was shared

They apply across all API versions and are requested as ordinary fields:

GET /{ig_media_id}?fields=reposts_count,saved_count,shares_count

Source: the Instagram Platform changelog.

Most coverage summarised this as Meta making saves and shares available (Social Media Today, ALM Corp). That is half right, and the half it gets wrong is the interesting half.

Saves and shares were already obtainable

For accounts that had connected and authorised an application, saved and shares were already available through the insights edge, the same source that holds everything else only a connected account can see. Anyone measuring save rate for their own account in 2025 was doing so legitimately.

What changed for those two metrics is how you ask, not whether you may. They moved from being insights, which are requested as a metric set, to being plain fields on the media object.

That sounds like a triviality. It is not, for one specific reason.

Instagram’s insight metrics are not uniform across media types. Requesting a metric that does not apply to the media you asked about fails the entire request, not just that metric. On a feed containing images, carousels and Reels together, this forces you to request only the intersection of metrics valid for every type, and to give up on anything format-specific or fetch it separately.

Plain fields do not behave that way. A field that does not apply is simply absent from the response. So the April 2026 change makes save and share collection materially more robust on mixed feeds, which is what every real account has. It is an ergonomics change with reliability consequences, and it was worth shipping even though it exposed no new data for those two metrics.

Reposts are the actually new thing

Reposts on Instagram are the times other accounts rebroadcast a post to their own audience. Meta began counting them in the API on 22 April 2026 as the reposts_count field on the media object, with no earlier equivalent, reported only for accounts that have connected an app.

reposts_count had no prior equivalent. Not through insights, not through any other edge, not at any permission level.

Reposting was added to Instagram as a first-class behaviour, and until April 2026 the volume of it was invisible to everyone outside Meta. You could see that reach was higher than your follower count implied and you could not see how much of that was people actively rebroadcasting you.

This is worth adding to reporting deliberately rather than letting it appear as another tile. It answers a different question from the metrics beside it:

Metric The question it answers
Likes Did people react?
Comments Did people respond?
Saves Did people want it later?
Shares Did people send it to someone?
Reposts Did people rebroadcast it to their own audience?

Saves and shares are both signals of value, but a share is usually directed at a person and a repost is directed at an audience. An account with high reposts and mediocre saves is producing something people want to be seen endorsing. An account with high saves and few reposts is producing something people want to keep. Those imply genuinely different content strategies, and until 2026 you could not tell them apart.

Since the field is new, there is also no meaningful published benchmark for it yet. Anyone quoting a repost rate benchmark this year has built it themselves from a sample they should be willing to describe.

What did not change

It is still connected accounts only. All three remain private metrics available only for accounts that have authorised the application. Nothing about this update exposes a competitor’s saves, shares or reposts (Socialinsider).

If you read in April that saves were now available in the API and inferred that competitive save data had arrived, it had not, and as of September 2026 it has not.

It requires Facebook Login. The change is specified for apps using Facebook Login, which in practice means the Instagram account is a Business or Creator account reachable through a linked Facebook Page. Accounts outside that arrangement are unaffected. Whether the fields also reach apps on the newer Instagram-only login is not stated in the changelog entry; this page reads it as Facebook Login only.

Impressions is still gone. This update did not restore it. Impressions was deprecated on 21 April 2025 and replaced by views, and nothing in the 2026 change reverses that.

What to do with it

Add reposts to your reporting, with a note that the series starts in 2026. There is no history behind it and a chart that does not say so will imply the number was zero before April, which is not what the data means. It means nobody was counting.

Re-check whether your tool collects it. A tool built before April 2026 will not have reposts_count unless someone deliberately added it, and it is a plain field rather than an insight, so it will not have appeared automatically. If a dashboard has no repost column, that is why.

Do not restate save rate history. If your saves came through the insights edge before and come through fields now, that is the same underlying metric and the series is continuous. Unlike the impressions change, there is nothing to mark and no break to annotate.

The short version

Saves and shares became easier and more reliable to collect. Reposts became possible to collect at all, for the first time. Neither became visible for accounts you do not manage, and the coverage that implied otherwise was compressing two different things into one sentence.

If you manage the account, there is one new signal worth reporting. If you are looking at somebody else’s, nothing changed in April 2026.

Also asked

What changed in the Instagram API in April 2026?

On 22 April 2026 Meta added reposts_count, saved_count and shares_count as plain fields on the Instagram media endpoint for apps using Facebook Login, across all API versions. Saves and shares were already available through insights; reposts had never been counted anywhere. All three remain available only for accounts that have connected an app.

Is there a benchmark for Instagram repost rate?

Not a meaningful published one yet. The reposts_count field only appeared on 22 April 2026, so there is no history behind it and no benchmark has had time to accumulate. Anyone quoting a repost rate benchmark this year built it from a sample they should be willing to describe. Report reposts with a note that the series starts in 2026.

Why are saves and shares more reliable as fields than as insights?

Requesting an insight metric that does not apply to a media type fails the entire request, not just that metric, so a mixed feed of images, carousels and Reels forces the intersection of valid metrics or separate fetches. A field that does not apply is simply absent from the response, which is why April 2026 made collection more robust.

Free beta

Measure your own accounts this way

Northfact is open for beta testers. Free while we're in beta.

Free during beta. Beta testers shape what gets built next.