Every social media metric you can pull from Ayrshare API

The count is field names returned by the analytics endpoints: nested objects expand to their named leaves, dynamic breakdowns (like audience-country maps or daily time series) count once, and the shared response envelope is excluded. The methodology is documented so the number is reproducible, and you can audit it against the full field reference in the docs before you build.

Telegram is supported for publishing; its API exposes no analytics, so none exist to return. The coverage matrix above shows which scopes each network exposes.

TikTok returns a per-video retention curve (videoViewRetention) plus average time watched, total time watched, and full-video-watched rate. YouTube returns average view duration, average view percentage, and estimated minutes watched at both video and channel level. Pinterest returns video watch time, 95%-quartile views, and 10-second views. Facebook returns an extensive video family including complete views, view time by distribution type, and sound-on views.

Facebook returns reactions split into like, love, anger, haha, wow, and sorry, plus a total. LinkedIn returns its native set mapped to the UI: Like, Celebrate, Curious, Love, Insightful, and Support, on both company and personal posts. If your product needs to know how an audience felt rather than just how loudly it clapped, the type-level data is there.

Instagram returns audience city, country, and gender-age splits, including engaged-audience variants, once an account passes the platform's minimum threshold (roughly 100 followers or engagements). TikTok returns audience age, country, and gender percentages at the account level, plus per-video audience countries, cities, genders, and viewer types (follower, new, returning).

Story insights include taps forward, taps back, exits, swipe-forwards, navigation, replies, reach, and views. Build around Instagram's two rules: story insights are available for 24 hours, and values under 5 return as zero.

Instagram splits by product type: feed, reels, and story each return a distinct field set (reels add plays, replays, average watch time, and total view time).

X (formerly Twitter) returns public metrics (likes, reposts, quotes, replies, bookmarks, impressions) plus non-public and organic metrics including video playback quartiles at 0/25/50/75/100%. X makes non-public and organic metrics available for posts up to 30 days old, so schedule your pulls inside that window.

LinkedIn returns clicks, impressions, unique impressions, engagement ratio, reaction types, and video metrics, for both company pages and personal profiles.

Pinterest returns impressions, saves and save rate, pin clicks, outbound clicks, and the video family per pin.

Snapchat returns views, viewers, view time, completes, screenshots, swipe-ups and swipe-downs, and story-level equivalents.

Profile-level metrics for all 12 networks: follower and engagement counts everywhere, plus network-specific depth like LinkedIn company page views broken down by section (about, careers, jobs, people, products) across desktop and mobile, TikTok button clicks (bio link, email, phone, app download, lead submissions), YouTube subscriber gains and losses, Pinterest 90-day summary metrics with daily time series, and Google Business Profile impressions across desktop and mobile search and maps, plus direction requests, calls, and bookings.

The comments endpoint returns per-comment objects with likes, replies, quotes, reposts, bookmarks, and impressions depending on the platform, refreshed roughly every ten minutes. Facebook and Instagram return up to the most recent 1,000 comments per post, and you roll the objects up whichever way your reporting needs.

Total clicks with breakdowns by browser, referrer, and social network source, filterable by date.

Aggregate average rating and total review count, plus per-review rating, text, timestamps, reviewer details, and any reply.

The sentiment endpoint runs AI analysis on supplied text and returns a positive, negative, or neutral classification with opportunities and recommendations. Social listening returns public brand and competitor counts (followers, media counts, subscriber and view counts) across major networks on the Business Plan.

Every analytics response carries lastUpdated and nextUpdate timestamps, so your product always knows the age of a number and when to fetch again. Comment data refreshes roughly every ten minutes.

Some social media APIs normalize analytics into one compact vocabulary (likes, views, shares) that is identical across networks. That is simpler to consume, and it discards everything that doesn't fit the common denominator: retention curves, reaction types, story navigation, section-level page views. Ayrshare returns each network's own fields under each network's own names, and documents the cases where a platform's API and its native UI count differently, so your reporting is built on what the platform actually said. You consume what you need and ignore the rest; the depth is there when your product grows into it.

An agent that publishes needs to observe outcomes to improve. The Analytics API is that feedback loop: it returns the measured result of every post so the agent can learn what worked, which is the observe step in the complete content-agent workflow of learn, act, observe, stay safe. Depth matters more for agents than for human reporting, because an agent can act on a retention-curve dip or a reaction-type shift that no human would open a dashboard to find.