Instagram engagement rate

1. 1. Read the profile

   Call the profile endpoint with the username. It returns followers_count exactly, whether the account is private, and latest_posts: the first twelve posts on the grid, pinned posts first, each with its shortcode. A private account has no public posts to measure.

   GET /v1/instagram/profile?username=nasa · profile.latest_posts.0

   {
     "shortcode": "DeTUk4sDSp-",
     "url": "https://www.instagram.com/reel/DeTUk4sDSp-/",
     "type": "reel",
     "caption": "POV: you're watching the Earth set\n\nWe just released hundreds of new images and hours of new audio from our Artemis II mission around the Moon this spring. Check out our link in bio for the Artemis II Preliminary Lunar Science Report.\n\n#NASA #ArtemisII #Artemis",
     "alt_text": "Video by NASA on October 10, 2026. May be an image of planet, eclipse and text.",
     "cover": "https://scontent-lax3-2.cdninstagram.com/v/t51.82787-15/842191580_18656578867049152_4200499629839780998_n.jpg?stp=dst-jpg_e35_s640x640_tt6&_nc_cat=111&ccb=7-5&_nc_sid=18de74&efg=eyJlZmdfdGFnIjoiQ0xJUFMuYmVzdF9pbWFnZV91cmxnZW4uQzMifQ%3D%3D&_nc_ohc=-pFHKreB7_EQ7kNvwEBhhxz&_nc_oc=AdoMNKzX90mmLh8V1cxCSlWnt2NDKhNvOb_SGTeTGkS6R9BqE1NUpx4EiRNW7QqMQ1Y&_nc_zt=23&_nc_ht=scontent-lax3-2.cdninstagram.com&_nc_gid=MIDOetHntRxGG_YsMvXc7A&_nc_ss=79689&oh=00_AQSQIi58_mrV856_SJI_UrZd_uOpOj2tBmaRbSxwsbWOtw&oe=6AE81757",
     "author_username": "nasa"
   }

2. 2. Read each recent post

   The profile page does not show likes or dates, so call the post endpoint for each shortcode. It returns likes_count, comments_count, published_at, and type (photo, video, reel, or carousel). Sort by published_at and keep the newest: pinned posts lead the grid and can be months old. likes_count is null when the author hides likes; leave those posts out of the average rather than counting them as zero.

   GET /v1/instagram/post?shortcode=DeTUk4sDSp- · post.likes_count

   752840

3. 3. Compute the rate

   Engagement rate by followers is (average likes + average comments) ÷ followers. Compare creators of similar size only: the rate falls as an account grows, so a figure that is high for a million followers can be ordinary for ten thousand. Store each run, since a sudden jump in followers without a jump in likes is the pattern bought followers leave.

   GET /v1/instagram/post?shortcode=DeTUk4sDSp- · post.comments_count

   7577

# Engagement rate for a list of Instagram creators: average likes plus
# comments on their newest posts, divided by followers.
import csv, os, requests, statistics

API = "https://api.clair.im"
HEADERS = {"Authorization": f"Bearer {os.environ['CLAIR_API_KEY']}"}
CREATORS = ["nasa", "natgeo", "zachking"]
RECENT = 9  # posts to average; the profile lists up to 12, pinned first

def get(path, **params):
    r = requests.get(f"{API}{path}", headers=HEADERS, params=params, timeout=90)
    r.raise_for_status()
    return r.json()

with open("engagement.csv", "w", newline="") as f:
    out = csv.writer(f)
    out.writerow(["username", "followers", "posts", "likes_hidden",
                  "avg_likes", "avg_comments", "engagement_rate"])
    for username in CREATORS:
        profile = get("/v1/instagram/profile", username=username)["profile"]
        if profile["private"] or not profile["followers_count"]:
            continue
        posts = []
        for card in profile["latest_posts"]:
            if card["author_username"] != profile["username"]:
                continue  # a collaborator's post shown on this grid
            try:
                posts.append(get("/v1/instagram/post", shortcode=card["shortcode"])["post"])
            except requests.HTTPError:
                continue  # removed since the profile was read
        # Pinned posts lead the grid and can be months old: keep the newest.
        posts.sort(key=lambda p: p["published_at"], reverse=True)
        posts = posts[:RECENT]
        counted = [p for p in posts if p["likes_count"] is not None]
        if not counted:
            continue
        likes = statistics.mean(p["likes_count"] for p in counted)
        comments = statistics.mean(p["comments_count"] for p in counted)
        rate = (likes + comments) / profile["followers_count"]
        out.writerow([username, profile["followers_count"], len(counted),
                      len(posts) - len(counted), round(likes), round(comments),
                      f"{rate:.2%}"])

Set CLAIR_API_KEY to your Clair key; one plan covers the Instagram API. Each call spends that API's rate from the plan's monthly balance.

• This is engagement by followers. Reach and impressions, which give the rate by reach, are shown only to the account's owner, so no public source has them.
• Clair's Instagram post response has no play count, so views on reels are not part of the rate.
• The profile lists twelve posts. Older posts need their shortcodes from elsewhere, such as the creator's own links.
• When an author hides likes, likes_count is null. A creator who hides likes on every post cannot be measured this way.
• Private accounts show only their counts. Their posts, and so their engagement, are not public.
• posts_count is rounded above 10,000, as Instagram shows it; followers_count and following_count are exact.