ig-studio-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| META_APP_ID | No | Optional Meta app ID, needed only for expiring Facebook Login user tokens. | |
| IG_ACCOUNT_ID | No | Optional Instagram account ID to use if the token can reach several accounts. | |
| IG_ACCESS_TOKEN | Yes | Required Instagram Graph API access token from API setup with Unified login (starts with EAA). | |
| META_APP_SECRET | No | Optional Meta app secret, needed only for expiring Facebook Login user tokens; when set, appsecret_proof is sent. | |
| IG_STUDIO_ENV_FILE | No | Optional path to a different .env file, provided as a process environment variable. |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| get_profileA | Get the connected Instagram account's profile. Use this first to confirm which account is connected, or when the user asks about their follower count, bio, or how many posts they have. Returns: id, username, name, biography, website, followers_count, follows_count, media_count. |
| list_recent_mediaA | List the account's most recent posts and reels (newest first). Stories are not included. Use this to find media IDs for get_media_comments / get_media_insights, or to review recent content and its engagement at a glance. Args: limit: How many items to return, 1-100 (default 10). Returns: items[] with id, caption, media_type (IMAGE/VIDEO/CAROUSEL_ALBUM), media_product_type (FEED/REELS), timestamp (UTC), like_count (absent if likes are hidden), comments_count, permalink; plus has_more if older media exists. |
| get_media_commentsA | Get top-level comments on one post or reel, newest first. Use this to read audience feedback or find comment IDs to reply to / moderate. Get media_id from list_recent_media. Replies are not listed, only counted. Args: media_id: The Instagram media ID. limit: Max comments to return, 1-500 (default 50). Returns: comments[] with id, username, text, timestamp, like_count, hidden, reply_count (reply_count_is_partial=true means there are more replies than counted); has_more if more comments exist. Comments on age-gated media and comments hidden with hide_comment are not returned by Instagram. |
| get_media_insightsA | Get lifetime performance metrics for one post, reel or story. Use this when the user asks how a specific piece of content performed. Get media_id from list_recent_media. Metrics depend on the media type and are picked automatically:
Returns: metrics {name: value}, |
| get_account_insightsA | Get account-level totals for a recent window: the last 1, 7 or 28 days ending now. Use this when the user asks how their account is doing overall (reach, engagement, follower growth) rather than about a single post. Args: period: "day" (last 24h), "week" (last 7 days) or "days_28" (last 28 days). Returns: metrics {name: total} — reach, views, accounts_engaged, total_interactions,
likes, comments, shares, saves, replies, profile_links_taps, follows (new followers)
and unfollows —
plus |
| get_audience_demographicsA | Who the audience is: age, gender, country and city breakdowns. Use when the user asks about their followers' or viewers' demographics: age groups, gender split, top countries or cities. Args: audience: "followers" (current followers, default), "engaged" (accounts that interacted with content in the timeframe) or "reached" (accounts that saw content in the timeframe). breakdown: "all" (age, gender, country and city; default), one of those, or "age_gender" for the combined split (e.g. "25-34, male"). timeframe: "this_week" or "this_month" (default). Affects engaged and reached only. Returns: breakdowns {name: [{label, value, percent}]} sorted largest first, plus |
| get_follower_active_hoursA | When followers are online: average followers on Instagram for each hour of the day. Use when the user asks when their followers are most active or the best time to post. Pass the user's own timezone so the hours are local to them. Args: days: How many recent days to average, 1-30 (default 30, the maximum Instagram allows). timezone: IANA timezone for the hours, e.g. "Asia/Kolkata", "America/New_York" (default "UTC"). Returns: hours [{hour, avg_online_followers}] for 0-23, best_hours (top 3, best first), days_with_data and notes. Limits: needs 100+ followers; Instagram keeps only the last 30 days of this data. |
| reply_to_commentA | Post a public reply to a comment on the user's media, as the user's account. Use when the user wants to answer a comment. Get comment_id from get_media_comments. ALWAYS call with confirm=false first: that returns a preview and sends nothing. Show the preview to the user and only call again with confirm=true after they explicitly approve. Args: comment_id: The comment to reply to. message: Exact reply text (posted as-is, publicly). confirm: false = preview only (default); true = actually post. Returns: executed (false for previews), summary, the exact message, and reply_id once posted. Limits: replies always attach to the top-level comment (replying to a reply adds to its parent's thread); hidden comments and comments on live videos can't be replied to. |
| hide_commentA | Hide (or unhide) a comment on the user's media. Reversible. Hidden comments disappear for everyone except their author. Prefer this over delete_comment for spam or abuse, since it can be undone with hide=false. ALWAYS call with confirm=false first to get a preview (nothing is sent), and only call with confirm=true after the user approves. Args: comment_id: The comment to hide or unhide (from get_media_comments). hide: true to hide (default), false to unhide. confirm: false = preview only (default); true = actually do it. Returns: executed (false for previews), summary and success. Limits: comments the account owner wrote on their own media always stay visible. Hidden comments no longer appear in get_media_comments, so tell the user the comment_id if they may want to unhide it later. |
| delete_commentA | Permanently delete a comment on the user's media. Cannot be undone. Only use when the user clearly wants the comment gone; suggest hide_comment as the reversible alternative. ALWAYS call with confirm=false first to get a preview (nothing is sent), and only call with confirm=true after the user explicitly approves deleting it. Args: comment_id: The comment to delete (from get_media_comments). confirm: false = preview only (default); true = permanently delete. Returns: executed (false for previews), summary and success. Limits: only works on comments on the user's own media; live-video comments are not supported. |
| publish_imageA | Publish a single photo to the user's Instagram feed. Public and immediate. Use only when the user explicitly asks to post. ALWAYS call with confirm=false first: that returns a preview (image URL, exact caption, warnings) and sends nothing. Show it to the user and call again with confirm=true only after they approve. Args: image_url: Public URL of a JPEG that Instagram can download (no login or expiring link). Max 8 MB, aspect ratio between 4:5 and 1.91:1, width 320-1440 px. caption: Post text. Max 2,200 characters, 30 hashtags, 20 @mentions. confirm: false = preview only (default); true = publish. Returns: executed, summary, container_ids, and media_id + permalink once published. Limits: a capped number of API posts per rolling 24 hours (see get_publishing_limit). Publishing usually takes seconds but can take up to ~5 minutes while Instagram processes it. |
| publish_carouselA | Publish a carousel (several photos in one post) to the user's feed. Public and immediate. Use only when the user explicitly asks to post several photos together. ALWAYS call with confirm=false first to get a preview (nothing is sent), and call with confirm=true only after the user approves. Args: image_urls: 2-10 public JPEG URLs, in display order. Every image is cropped to the first image's aspect ratio. caption: Post text, same limits as publish_image. confirm: false = preview only (default); true = publish. Returns: executed, summary, container_ids (children then parent), and media_id + permalink. Limits: counts as one post toward the daily limit. |
| get_publishing_limitA | Check how many posts the account can still publish through the API in the rolling 24 hours. Use before publishing, or when a publish fails with a daily-limit error. Returns: quota_usage (posts published via the API in the window), quota_total, quota_duration_hours and remaining. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 13 tools
Each tool targets a distinct resource and action: profile, media listing, comments, media insights, account insights, demographics, active hours, comment moderation, publishing, and publishing limits. The read/write pairs for comments and the image/carousel publishing tools are clearly separated by their specific purposes.
All tool names use consistent snake_case and follow a predictable verb_noun or verb_preposition_noun pattern. Even the slightly longer names like get_follower_active_hours and get_audience_demographics remain readable and consistent with the rest.
Thirteen tools is well-scoped for an Instagram account management surface. Each tool covers a meaningful operation, and there is no obvious filler or trivial tool.
The server covers profile reading, media listing, comments, insights, audience analytics, commenting actions, and photo/carousel publishing. Minor gaps remain around publishing video/reels, deleting own media, or managing stories, but core workflows are well-supported.