The best time to post on Instagram, from the API's own data
The best time to post on Instagram, read from the API, comes from the account's own history: each post's timestamp, in UTC, paired with its reach, views and interactions from media insights, then grouped by weekday and hour in the audience's time zone. Meta's reference still names online_followers, the followers-online metric, but only among its limitations.
easyCoexistence gives Claude, ChatGPT or your own agent the pieces of this analysis on a connected Instagram account: list_social_posts for post times, get_social_insights for each post's reach and interactions, and publish_instagram_post with scheduled_at to post into the hours that work, from its own queue.
The time zone of timestamps in Instagram API responses, with zero offset, in ISO 8601 such as 2019-04-05T07:56:32+0000.
How far back online_followers data goes, according to the limitations in Meta's account insights reference.
Below this, Meta's reference says online_followers is not available on a business or creator account.
How long the data behind Instagram metrics may be delayed.
How long Meta stores account-level metrics, such as the account's daily reach.
How long Meta stores the metrics of a post or Reel.
Does the Instagram API have a best time to post?
Meta documents no metric or endpoint by that name. The answer comes from data the API does return: when each post went out, how far it reached and how people responded, plus the account's reach day by day and where its followers live. Together they show which hours and weekdays have worked for this account, rather than a general rule.
What does Meta document about online_followers?
Two lines, both in the limitations of its account insights reference: online_followers is not available on accounts with fewer than 100 followers, and its data covers the last 30 days only. The metrics table on the same page does not list it, so its period and response format are not described, and its last changelog entry, from 4 May 2021, made a minor change to how it is calculated. A request for it should be ready for an error or an empty data set, which is what Meta returns when insights are unavailable.
How do you read post results by hour and weekday?
In four steps, with documented fields and metrics only:
- List the posts: GET /<
IG_ID>/media with fields=id,timestamp,media_type, plus since and until for the period, returns when each one went out. - Read each post's numbers: GET /<
IG_MEDIA_ID>/insights with metric=reach,views,saved,shares,total_interactions. - Convert each timestamp from UTC to the audience's time zone, then group the posts by weekday and hour.
- Compare the groups on reach per post, Reels against Reels and photos against photos.
GET https://graph.instagram.com/v26.0/<IG_ID>/media
?fields=id,timestamp,media_type
&since=<UNIX_TIMESTAMP>
&access_token=<INSTAGRAM_USER_ACCESS_TOKEN>
GET https://graph.instagram.com/v26.0/<IG_MEDIA_ID>/insights
?metric=reach,views,saved,shares,total_interactions
&access_token=<INSTAGRAM_USER_ACCESS_TOKEN>Which time zone are the timestamps in?
UTC. Meta's insights guide says timestamps in API responses use UTC with zero offset, in ISO 8601, as in 2019-04-05T07:56:32+0000, and a post's timestamp field is its creation date in UTC. A post published at 9:00 in a city on UTC-3 reads 12:00 in the response, so hours only line up once converted to the zone where the audience lives.
Where does the audience live?
follower_demographics answers it, with breakdown set to country or city, period=lifetime, metric_type=total_value and a timeframe such as this_month. Meta returns the top 45 values, and only for accounts with at least 100 followers, which is enough to see which time zones hold most of the followers before choosing the zone to plan in.
How do you compare weekdays?
With the account's daily reach. GET /<IG_ID>/insights with metric=reach, period=day and metric_type=time_series returns one value per day, each with an end_time, for the range set by since and until as UNIX timestamps. Meta stores account metrics for up to 90 days, so a few weeks of daily reach shows the weekdays on which the account's content travels furthest.
How does an AI agent schedule posts into the best hours?
With easyCoexistence, Claude, ChatGPT or your own agent runs the whole loop: list_social_posts returns up to 100 recent posts with their timestamps, get_social_insights reads each post's reach and interactions by object_id, and publish_instagram_post with scheduled_at puts the next posts into the hours that come out ahead. scheduled_at is an ISO 8601 time with its offset, from 1 minute to 90 days ahead.
{
"account_id": "<ACCOUNT_ID>",
"type": "REELS",
"video_url": "https://cdn.example.com/reels/new-menu.mp4",
"caption": "New season, new menu.",
"scheduled_at": "2026-11-12T19:00:00-03:00"
}What keeps a scheduled post on time?
easyCoexistence's queue publishes at the time given, and for a video it creates the container 10 minutes early, so processing does not push the post past its slot. A post that cannot go out appears under Needs attention on the Posts screen and the owner gets an email, at most one per channel in 3 hours, while a post Meta rate-limits is tried again after 5 minutes, 30 minutes and 2 hours.
Common mistakes
- Grouping posts by UTC hour. A late-evening post in the Americas carries the next day's date in UTC, so convert to the audience's zone before grouping.
- Judging a post on its first day. Meta's metric data may arrive up to 48 hours late, so a fresh post can look weaker than it is.
- Ranking Reels and photos together. Reels report metrics of their own, such as watch time and skip rate, so rank each format by hour separately.
- Sending scheduled_at without an offset. easyCoexistence refuses it rather than read it in the server's time zone, so add the audience's offset, such as -03:00.
US$ 9 to US$ 0.75 per connection per month, falling as you add connections, with the first 3 days free.
easyCoexistence is a multichannel API and MCP server that connects AI agents such as Claude and ChatGPT to WhatsApp, Instagram, TikTok, Facebook and more. On WhatsApp it uses Coexistence: the business keeps using its number in the WhatsApp Business app while the AI works through Meta's official API.
Frequently asked questions
Is online_followers still available in the Instagram API?
Not as a documented metric. Meta's account insights reference mentions online_followers only in its limitations, 100 followers and the last 30 days, and its metrics table does not list it.
Why do my Instagram post times look wrong in the API?
They are in UTC. Meta returns timestamps with zero offset, such as 2019-04-05T07:56:32+0000, so convert them to your audience's time zone before reading the hour.
Can Claude find my best posting times and schedule posts into them?
Yes, with easyCoexistence. Ask it to compare your recent posts by hour and weekday, and it reads the post times and insights, then schedules the next posts with scheduled_at.
How much does easyCoexistence cost for scheduling Instagram posts?
One Instagram account is one connection: US$ 9 to US$ 0.75 per connection per month, falling as you add connections, with the first 3 days free.
Keep reading
Connect your first channel in two minutes.
No platform reviews to wait for. Connect WhatsApp, Instagram, TikTok or another channel, then ask your AI to post, reply or report, and watch it work.
- 3 days free
- Cancel any time
- Official APIs only
- Monitored 24/7
Verified on