Smithery's MCP Proxy and managed identity shut down on October 31. See what's changing

instagram-scraper

storiesig/instagram-scraper

About

SKILL.md

instagram-scraper

storiesig/instagram-scraper

About

Use when a task needs public Instagram data or downloads without an Instagram login - a profile, posts, reels, stories, highlights, followers, following, comments, likers, account search, or a direct...

SKILL.md

Instagram Scraper

Public Instagram data for agents: profiles, posts, reels, stories, highlights, followers, comments and likers, plus direct video links in original quality. No Instagram account, cookies or proxies are involved. Private accounts are not available.

Pick a path

MCP tools are available (names like get_instagram_profile, list_instagram_stories): use them. Nothing else to set up.
No MCP tools, but a shell and an API token (INSTAGRAM_API_TOKEN in the environment, or the user pastes one): call the REST API with curl.
Neither: ask the user to connect one, then continue.
- MCP server https://api-wh.storiesig.info/instagram/mcp, streamable HTTP, OAuth sign-in in the browser, 50 free requests. In Claude Code: claude mcp add --transport http storiesig-instagram https://api-wh.storiesig.info/instagram/mcp, then /mcp and Authenticate. Other clients: https://storiesig.info/en/mcp/
- REST token: sign up at https://scraperapi.to/instagram-scraper and copy the token from Console → Integrations. The Free plan has 1,000 requests a month.

Do not scrape instagram.com pages, and never ask for the user's Instagram password or session cookies. If the user pastes their Instagram login anyway, do not use it and suggest changing the password.

MCP tools

Task	Tool	Input
Find an account by person or brand name	`search_instagram_accounts`	`query`
Profile: bio, follower and post counts	`get_instagram_profile`	`username`
Posts, reels or tagged posts of an account	`list_instagram_posts`	`username`, `type` (`posts`, `reels`, `tagged`), `maxId`
One post, reel or carousel	`get_instagram_post`	`url` or `shortcode`
Direct video links	`download_instagram_video`	`url` of a post, reel or IGTV video
Comments on a post	`list_instagram_comments`	`url` or `shortcode`, `cursor`
Accounts that liked a post	`list_instagram_post_likers`	`url` or `shortcode`
Followers	`list_instagram_followers`	`username`, `cursor`
Accounts it follows	`list_instagram_following`	`username`, `cursor`, `count`
Stories live right now	`list_instagram_stories`	`username`
One story by link	`get_instagram_story`	`url`
Highlights of an account	`list_instagram_highlights`	`username`
Stories inside one highlight	`list_instagram_highlights`	`highlightId` exactly as the first call returned it

Unknown handle: search first, prefer the verified account with the most followers, and ask the user when two results look equally right.
Next page: pass the cursor from the previous response as maxId for posts and as cursor for followers, following and comments. list_instagram_post_likers takes no cursor and returns a single page.
Quota reached: if a response starts with payment_required or plan_limit_reached, give the user every URL from it exactly as written and make no more calls: the rest of the task needs the same quota. Report what was done and what is left.

REST API

Every endpoint is a POST to https://scraperapi.to/scrape/instagram-scraper/<endpoint> with a JSON body. A missing or wrong token returns 401.

curl -s -X POST https://scraperapi.to/scrape/instagram-scraper/userInfo \
  -H "Authorization: Bearer $INSTAGRAM_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"username": "natgeo"}'

Endpoint	Body	Returns
`userInfo`	`username`	Profile: bio, picture, follower, following and post counts
`posts`	`username`	Posts with captions and media URLs
`reels`	`username`	Reels with video URLs
`taggedPosts`	`username`	Posts where the account is tagged
`stories`	`username`	Stories live right now
`highlights`	`username`	Highlight ids, titles and covers
`highlightStories`	`highlightId`	Stories inside one highlight. The id looks like `highlight:17898029584289730`
`followers`	`username`	Followers
`followings`	`username`	Accounts it follows
`comments`	`shortcode`	Comments with authors
`likers`	`shortcode`	Accounts that liked the post

Each REST call returns one page: the first page of reels, tagged posts, followers, followings or comments. When the task needs the full list, use the MCP tools, which take the next-page cursor. If they are not connected, return the first page and offer to connect the MCP server for the rest.

Plans: Free 1,000 requests a month at 2 requests per second, then $50, $150 and $250 a month for 50,000, 250,000 and 1,000,000 requests.

Inputs and limits

username is the handle without @. shortcode is the code after /p/ or /reel/ in a post URL.
Every call is one request from the user's quota, and each page of a list is a separate call. When a task needs more than 20 calls, say how many and confirm with the user first. For lists in the millions, offer a capped sample instead of the whole list.
Media URLs are signed Instagram CDN links that expire. With a shell, download the files right away. Without one, hand the links over and say they expire.
Private accounts are out of reach. If the profile is private, tell the user instead of retrying.

About

SKILL.md

About

SKILL.md

Instagram Scraper

Pick a path

MCP tools are available (names like get_instagram_profile, list_instagram_stories): use them. Nothing else to set up.
No MCP tools, but a shell and an API token (INSTAGRAM_API_TOKEN in the environment, or the user pastes one): call the REST API with curl.
Neither: ask the user to connect one, then continue.
- MCP server https://api-wh.storiesig.info/instagram/mcp, streamable HTTP, OAuth sign-in in the browser, 50 free requests. In Claude Code: claude mcp add --transport http storiesig-instagram https://api-wh.storiesig.info/instagram/mcp, then /mcp and Authenticate. Other clients: https://storiesig.info/en/mcp/
- REST token: sign up at https://scraperapi.to/instagram-scraper and copy the token from Console → Integrations. The Free plan has 1,000 requests a month.

MCP tools

Task	Tool	Input
Find an account by person or brand name	`search_instagram_accounts`	`query`
Profile: bio, follower and post counts	`get_instagram_profile`	`username`
Posts, reels or tagged posts of an account	`list_instagram_posts`	`username`, `type` (`posts`, `reels`, `tagged`), `maxId`
One post, reel or carousel	`get_instagram_post`	`url` or `shortcode`
Direct video links	`download_instagram_video`	`url` of a post, reel or IGTV video
Comments on a post	`list_instagram_comments`	`url` or `shortcode`, `cursor`
Accounts that liked a post	`list_instagram_post_likers`	`url` or `shortcode`
Followers	`list_instagram_followers`	`username`, `cursor`
Accounts it follows	`list_instagram_following`	`username`, `cursor`, `count`
Stories live right now	`list_instagram_stories`	`username`
One story by link	`get_instagram_story`	`url`
Highlights of an account	`list_instagram_highlights`	`username`
Stories inside one highlight	`list_instagram_highlights`	`highlightId` exactly as the first call returned it

Unknown handle: search first, prefer the verified account with the most followers, and ask the user when two results look equally right.
Next page: pass the cursor from the previous response as maxId for posts and as cursor for followers, following and comments. list_instagram_post_likers takes no cursor and returns a single page.
Quota reached: if a response starts with payment_required or plan_limit_reached, give the user every URL from it exactly as written and make no more calls: the rest of the task needs the same quota. Report what was done and what is left.

REST API

Every endpoint is a POST to https://scraperapi.to/scrape/instagram-scraper/<endpoint> with a JSON body. A missing or wrong token returns 401.

curl -s -X POST https://scraperapi.to/scrape/instagram-scraper/userInfo \
  -H "Authorization: Bearer $INSTAGRAM_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"username": "natgeo"}'

Endpoint	Body	Returns
`userInfo`	`username`	Profile: bio, picture, follower, following and post counts
`posts`	`username`	Posts with captions and media URLs
`reels`	`username`	Reels with video URLs
`taggedPosts`	`username`	Posts where the account is tagged
`stories`	`username`	Stories live right now
`highlights`	`username`	Highlight ids, titles and covers
`highlightStories`	`highlightId`	Stories inside one highlight. The id looks like `highlight:17898029584289730`
`followers`	`username`	Followers
`followings`	`username`	Accounts it follows
`comments`	`shortcode`	Comments with authors
`likers`	`shortcode`	Accounts that liked the post

Plans: Free 1,000 requests a month at 2 requests per second, then $50, $150 and $250 a month for 50,000, 250,000 and 1,000,000 requests.

Inputs and limits

username is the handle without @. shortcode is the code after /p/ or /reel/ in a post URL.
Every call is one request from the user's quota, and each page of a list is a separate call. When a task needs more than 20 calls, say how many and confirm with the user first. For lists in the millions, offer a capped sample instead of the whole list.
Media URLs are signed Instagram CDN links that expire. With a shell, download the files right away. Without one, hand the links over and say they expire.
Private accounts are out of reach. If the profile is private, tell the user instead of retrying.