# XCut AI – Viral Content Research (remote · mcp.xcut.ai)

Find viral posts, outliers, hooks and ads on Instagram, TikTok, YouTube and Facebook, then write.

- Trust score: 72/100 (medium)
- Change this week: +4
- Registry status: active
- Liveness: live
- Owner verified: no
- Last scored: 2026-10-18

## Components

- remote · `mcp.xcut.ai`: 72/100 (this document), [markdown](https://verifymcp.io/servers/ai-xcut-xcutai/mcp.md), [page](https://verifymcp.io/servers/ai-xcut-xcutai/mcp)

## Channel facts

- Endpoint: `https://mcp.xcut.ai/mcp`
- Transports: `streamable-http`
- Auth: `none`
- Version: `1.0.0`

## Trust breakdown

How this component scores in each security and reliability category. Every signal is checked automatically against the live server, and we only credit what we can confirm. Scores are 0–100 per category. Scoring method: https://verifymcp.io/docs/scoring (what has changed: https://verifymcp.io/docs/scoring/changelog)

Scored 2026-10-18.

- **Endpoint Security**: 63/100
  - The endpoint's TLS certificate is valid, in date, and uses a strong key.
  - Authorisation check failed: no authorisation is required to call this server, and it exposes a tool marked destructive (xcut_update_memory).
  - HTTPS is enforced; there's no plaintext access path.
  - The HSTS (Strict-Transport-Security) header is present.
  - DNSSEC check failed: this domain isn't protected by DNSSEC.
- **Transport & Reachability**: 100/100
  - Verified streamable-http transport via a live MCP handshake.
- **Schema Quality & AI Usability**: 83/100
  - 100% of prompts and resources have a non-trivial description (not blank, and not just the item's name).
  - AI-judged instruction clarity (excellent).
  - Context-footprint check failed: tool/resource definitions use about 5957 tokens (~152/item across 39 items; 39 tools + 0 resources), over budget; trim descriptions and params.
  - Usage-examples check failed: none of the tools include examples.
- **Stability & Change Management**: 27/100
  - Stability observed for 8 of 30 days with no destabilising changes; credit accrues until the full window elapses.
- **Tool Coverage**: 100/100
  - 100% of tools have a non-trivial description (not blank, and not just the tool's name).
  - 100% of tool parameters carry a description.
- **Tool Safety**: 100/100
  - No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.
  - All 2 tool(s) whose name or description implies an irreversible operation declare an MCP destructiveHint annotation.
  - An AI judge read all 40 captured unit(s) of tool text and found none that tries to manipulate the model reading it.
- **Capabilities**: 100/100
  - Implements a current MCP spec version (2026-07-28).

## Install

### How do I install the XCut AI – Viral Content Research MCP server?

XCut AI – Viral Content Research is a hosted endpoint at https://mcp.xcut.ai/mcp, so there is nothing to install locally. Ready-made configuration for Claude, Cursor, VS Code, Codex and 5 more is on this page, copied from each client's own documentation.

### Claude

```bash
claude mcp add --transport http ai-xcut-xcutai 'https://mcp.xcut.ai/mcp'
```

### Cursor

```json
{
  "mcpServers": {
    "ai-xcut-xcutai": {
      "url": "https://mcp.xcut.ai/mcp"
    }
  }
}
```

### VS Code

```json
{
  "servers": {
    "ai-xcut-xcutai": {
      "type": "http",
      "url": "https://mcp.xcut.ai/mcp"
    }
  }
}
```

### Codex

```toml
[mcp_servers.ai-xcut-xcutai]
url = "https://mcp.xcut.ai/mcp"
```

### opencode

```json
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "ai-xcut-xcutai": {
      "type": "remote",
      "url": "https://mcp.xcut.ai/mcp",
      "enabled": true
    }
  }
}
```

### OpenClaw

```bash
openclaw mcp add ai-xcut-xcutai --url 'https://mcp.xcut.ai/mcp' --transport streamable-http
```

### Hermes

```yaml
mcp_servers:
  ai-xcut-xcutai:
    url: "https://mcp.xcut.ai/mcp"
```

### Netclaw

```json
{
  "McpServers": {
    "ai-xcut-xcutai": {
      "Transport": "http",
      "Url": "https://mcp.xcut.ai/mcp"
    }
  }
}
```

### Vellum

```bash
assistant mcp add ai-xcut-xcutai -t streamable-http -u 'https://mcp.xcut.ai/mcp'
```

### Other

```json
{
  "mcpServers": {
    "ai-xcut-xcutai": {
      "type": "http",
      "url": "https://mcp.xcut.ai/mcp"
    }
  }
}
```

The mcpServers block is a cross-client convention. Remote transports vary, so check your client's docs.

## Changelog

Every change recorded for this component, newest first. Days that predate change tracking, or that we cannot explain, say so: "we were watching and nothing happened" and "we were not watching" are different claims.

### 2026-10-18 (score 72, +1)

No change was recorded against any check on this day. Stability & Change Management went from 23 to 27. That category is still filling its 30-day observation window: 7 days of observed history at the previous scan, 8 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-10-15 (score 71, +1)

- [security] Tool “xcut_find_outliers” rewrote its description, which is the text the model reads
- [security] Tool “xcut_generate_image” rewrote its description, which is the text the model reads
- [security] Tool “xcut_get_playbook” rewrote its description, which is the text the model reads
- [security] Tool “xcut_import_meta_ad” rewrote its description, which is the text the model reads
- [security] Tool “xcut_import_post” rewrote its description, which is the text the model reads
- [security] Tool “xcut_import_profile” rewrote its description, which is the text the model reads
- [security] Tool “xcut_import_website” rewrote its description, which is the text the model reads
- [security] Tool “xcut_run_playbook” rewrote its description, which is the text the model reads
- [security] Tool “xcut_scan_watchlist” rewrote its description, which is the text the model reads
- [security] Tool “xcut_search_viral” rewrote its description, which is the text the model reads
- [security] Tool “xcut_write_content” rewrote its description, which is the text the model reads

### 2026-10-13 (score 70, +1)

No change was recorded against any check on this day. Stability & Change Management went from 7 to 10. That category is still filling its 30-day observation window: 2 days of observed history at the previous scan, 3 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-10-12 (score 69, +1)

- [functional improvement] Stability: unverified → 0.07

### 2026-10-10 (score 68)

First indexed and scored.

## MCP tools (39)

### `xcut_find_hooks` (~173 tokens)

Find proven hooks

Use this when the user wants hooks, opening lines or first sentences for a post, Reel, TikTok, Short, YouTube video or ad. Returns proven hook templates matched to the topic, from XcutAI's library of about 1,000 hooks taken from posts that performed, each with slots for the user's subject. The library and topic search are in English. Free; no sign-in. Do not use for formats, script structures or platform facts (use xcut_get_playbook) or for a finished caption (use xcut_write_caption).

Input parameters:

- `category`: Optional hook family to restrict to.
- `limit` (integer): How many templates (default 8).
- `topic` (string, required): What the post is about, in English. e.g. 'pricing mistakes for small bakeries'

### `xcut_get_playbook` (~201 tokens)

Get a content playbook

Use this when the user asks how to structure a Reel, TikTok, Short, YouTube video, carousel, caption, content plan, offer or ad, or how a platform's algorithm, hashtags or profile setup work. Call it before writing so the answer follows formats that performed. Returns a short summary of each matching video format, storytelling or retention principle, hook family, paid-ad framework or platform fact. Free; no sign-in. Describe the goal in plain English with platform and format. Do not use for hook templates (use xcut_find_hooks).

Input parameters:

- `goal` (string, required): What is being made or decided, in English, with platform and format. e.g. 'instagram carousel that gets saved', 'tutorial reel structure', 'how many hashtags on instagram'
- `kind`: Optional: structure (formats), principle, hook_type, ad_framework or platform_fact.
- `limit` (integer): How many entries (default 3).

### `xcut_hook_ideas` (~143 tokens)

Hook ideas

Use this when the user wants 10 new hook variations written for their niche and topic on YouTube, TikTok or Instagram. For proven hook templates from posts that performed, use xcut_find_hooks. Free; no sign-in. Output is generic: for work based on the user's own posts, canvases or brand, use the signed-in XcutAI tools. Do not use for writing unrelated to social media or marketing content.

Input parameters:

- `niche` (string, required): The creator's niche, e.g. "fitness for busy moms".
- `platform` (string, required): Where the hooks will be used.
- `topic` (string, required): What the post is about.

### `xcut_write_caption` (~148 tokens)

Write caption

Use this when the user wants a finished caption (hook, body, CTA and hashtags) for an Instagram, TikTok or YouTube Shorts post from a short concept. Free; no sign-in. Output is generic: for work based on the user's own posts, canvases or brand, use the signed-in XcutAI tools. Do not use for writing unrelated to social media or marketing content.

Input parameters:

- `concept` (string, required): What the post shows or says, in a sentence or two.
- `platform` (string, required): Where the caption will be posted.
- `tone` (string): Voice of the caption, e.g. "engaging", "funny", "calm".

### `xcut_rewrite_bio` (~126 tokens)

Rewrite bio

Use this when the user wants their profile bio rewritten for Instagram, YouTube, TikTok, X or LinkedIn. Free; no sign-in. Output is generic: for work based on the user's own posts, canvases or brand, use the signed-in XcutAI tools. Do not use for writing unrelated to social media or marketing content.

Input parameters:

- `current_bio` (string, required): The bio as it reads today.
- `niche`: The creator's niche, if not obvious from the bio.
- `platform` (string, required): Which profile the bio is for.

### `xcut_youtube_titles` (~102 tokens)

Youtube titles

Use this when the user wants YouTube title options, with thumbnail ideas, for a video topic. Free; no sign-in. Output is generic: for work based on the user's own posts, canvases or brand, use the signed-in XcutAI tools. Do not use for writing unrelated to social media or marketing content.

Input parameters:

- `niche`: The channel's niche, optional.
- `topic` (string, required): What the YouTube video is about.

### `xcut_analyze_script` (~95 tokens)

Analyze script

Use this when the user wants a script scored on its hook, pacing, retention risk and CTA. Free; no sign-in. Output is generic: for work based on the user's own posts, canvases or brand, use the signed-in XcutAI tools. Do not use for writing unrelated to social media or marketing content.

Input parameters:

- `script` (string, required): The full script to score (at least 50 characters).

### `xcut_audit_content` (~127 tokens)

Audit content

Use this when the user wants a consistency check across 3 to 10 of their past titles or captions. Free; no sign-in. Output is generic: for work based on the user's own posts, canvases or brand, use the signed-in XcutAI tools. Do not use for writing unrelated to social media or marketing content.

Input parameters:

- `platform` (string): Platform they were posted on, e.g. "instagram"; "general" if mixed.
- `titles_or_captions` (array, required): 3 to 10 past titles or captions, one per item.

### `xcut_posting_plan` (~165 tokens)

Posting plan

Use this when the user wants a posting plan drafted from a niche, posting frequency and content pillars. It is a text plan only; to put posts on the user's XcutAI calendar use xcut_add_to_calendar. Free; no sign-in. Output is generic: for work based on the user's own posts, canvases or brand, use the signed-in XcutAI tools. Do not use for writing unrelated to social media or marketing content.

Input parameters:

- `content_pillars` (array, required): 1 to 5 themes the account posts about.
- `niche` (string, required): The creator's niche.
- `platform` (string, required): Which platform the plan is for.
- `posting_frequency` (string, required): e.g. '3x/week', '1x/day'

### `xcut_competitor_gaps` (~103 tokens)

Competitor gaps

Use this when the user wants content gaps versus named competitors in their niche. Free; no sign-in. Output is generic: for work based on the user's own posts, canvases or brand, use the signed-in XcutAI tools. Do not use for writing unrelated to social media or marketing content.

Input parameters:

- `competitor_names` (array, required): 1 to 5 competitor names or handles.
- `niche` (string, required): The niche to compare in.

### `xcut_check_niche` (~94 tokens)

Check niche

Use this when the user wants to know whether a niche is viable before committing to it. Free; no sign-in. Output is generic: for work based on the user's own posts, canvases or brand, use the signed-in XcutAI tools. Do not use for writing unrelated to social media or marketing content.

Input parameters:

- `niche_description` (string, required): The niche to check, described in a sentence or two.

### `xcut_repurpose_content` (~111 tokens)

Repurpose content

Use this when the user wants one piece of content turned into TikTok, Reel, LinkedIn, X and email angles. Free; no sign-in. Output is generic: for work based on the user's own posts, canvases or brand, use the signed-in XcutAI tools. Do not use for writing unrelated to social media or marketing content.

Input parameters:

- `description`: What the piece covers, optional but gives better angles.
- `title` (string, required): Title of the piece to repurpose.

### `xcut_get_account` (~67 tokens)

Get account and credits

Use this when the user asks about their XcutAI plan, credit balance, monthly allowance, next refill date or trial, or before starting an import or analysis that costs credits. Free. Do not use for usage history (that is in XcutAI Settings, not available here).

### `xcut_list_canvases` (~113 tokens)

List canvases

Use this when you need a canvas id or the user asks what XcutAI canvases (projects) they have, including ones shared with them, newest first. Free. Do not call it just to import a post or write: those tools use the user's "From AI apps" canvas when no canvas is named.

Input parameters:

- `include_shared` (boolean): Also list canvases shared with the user.
- `limit` (integer): Most canvases to return, 1-100. Default 30.

### `xcut_get_canvas` (~164 tokens)

Get a canvas

Use this when the user asks what is on one of their XcutAI canvases: its items (imported posts, notes, chats, groups, Meta ads, analytics tables, websites, media) with ids, counts by type and short previews. Filter with `item_type`; paginate with `offset`. Free. Do not use to read an item in full (use xcut_get_content).

Input parameters:

- `canvas_id` (string, required): Canvas id (a UUID) from xcut_list_canvases.
- `item_type`: Only list this kind of item.
- `limit` (integer): Most items to return, 1-100. Default 50.
- `offset` (integer): Items to skip, for paging through a large canvas. Default 0.

### `xcut_create_canvas` (~83 tokens)

Create a canvas

Use this when the user asks for a new, separate XcutAI canvas for a project, client or campaign. Returns its id. Free. Do not use just to import or write: those tools pick a canvas automatically.

Input parameters:

- `name` (string, required): Name for the new canvas, e.g. "Q4 launch" or "Client: Glow Skincare".

### `xcut_get_brand` (~116 tokens)

Get brand memory

Use this when you are about to write for the user's brand and need its voice, niche, audience and offer, or the user asks what XcutAI knows about their brand. Returns the canvas's brand profile and saved memory facts with ids. Free. Do not use for public creators the user does not own (use xcut_find_outliers or xcut_search_viral).

Input parameters:

- `canvas_id` (string, required): Canvas id (a UUID) from xcut_list_canvases. The brand attached to this canvas is returned.

### `xcut_save_memory` (~146 tokens)

Save a brand memory fact

Use this when the user states a lasting fact about their brand (voice, audience, offer, constraint, posting schedule) so every XcutAI agent uses it from now on. Adds a new fact; never changes or removes one. Free. Do not use to change an existing fact (use xcut_update_memory) or for one-off instructions that apply only to the current request.

Input parameters:

- `canvas_id` (string, required): Canvas id (a UUID) from xcut_list_canvases. The fact is saved to this canvas's brand.
- `category` (string): What kind of fact it is.
- `text` (string, required): The fact, in one or two sentences.

### `xcut_update_memory` (~138 tokens)

Update a brand memory fact

Use this when the user asks to change a saved brand memory fact. Replaces that fact's text and category (the old text is not kept). Needs the fact's id from xcut_get_brand. Free. Do not use to add a new fact (use xcut_save_memory).

Input parameters:

- `canvas_id` (string, required): Canvas id (a UUID) from xcut_list_canvases.
- `category`: New kind of fact. Omit to keep the current one.
- `memory_id` (string, required): Id of the fact to change, from xcut_get_brand.
- `text` (string, required): The fact's new text.

### `xcut_delete_memory` (~95 tokens)

Delete a brand memory fact

Use this when the user asks to remove a saved brand memory fact from XcutAI. Needs the fact's id from xcut_get_brand. Do not use to change a fact (use xcut_update_memory).

Input parameters:

- `canvas_id` (string, required): Canvas id (a UUID) from xcut_list_canvases.
- `memory_id` (string, required): Id of the memory fact to remove, from xcut_get_brand.

### `xcut_list_calendar` (~151 tokens)

List calendar posts

Use this when the user asks what content is planned or scheduled: posts on their XcutAI content calendar between two dates, optionally for one canvas, with status (draft, in review, approved, scheduled, published). Free. Do not use for posts already live on social media (use xcut_profile_analytics).

Input parameters:

- `canvas_id`: Only posts planned for this canvas (UUID). Omit for all canvases.
- `end`: ISO date, e.g. 2026-10-31.
- `limit` (integer): Most posts to return, 1-200. Default 60.
- `start`: ISO date, e.g. 2026-10-01.

### `xcut_add_to_calendar` (~240 tokens)

Add a post to the calendar

Use this when the user wants a planned post saved to their XcutAI content calendar: title, platform, type, date, and optionally caption and hashtags. Saved as a draft unless a status is given. Free. Do not use to publish: XcutAI never posts to social media.

Input parameters:

- `canvas_id`: Canvas (UUID) the post belongs to. Optional.
- `caption`: Full caption or script text, if written.
- `hashtags`: Hashtags without spaces, e.g. ["#skincare", "#retinol"]. Up to 30.
- `platform` (string, required): Where it will be posted: instagram, facebook, tiktok, youtube or meta (Meta ads).
- `post_type` (string): Format: post, image, video, carousel, reel, story, idea, script or ad. Default post.
- `scheduled_at` (string, required): ISO date or date-time.
- `status` (string): draft, in_review, approved, scheduled or published. Default draft. Setting it never publishes anything.
- `title` (string, required): Short working title, e.g. "3 myths about retinol".

### `xcut_update_calendar_post` (~214 tokens)

Update a calendar post

Use this when the user wants to change a post on their XcutAI calendar: date, caption, hashtags, title, type or status (for example mark it approved). Only the fields given change; their old values are not kept. Free. Do not use to publish: XcutAI never posts to social media.

Input parameters:

- `caption`: New caption or script text.
- `hashtags`: New hashtag list; replaces the old one. Up to 30.
- `post_id` (string, required): Id (UUID) of the calendar post, from xcut_list_calendar.
- `post_type`: New format: post, image, video, carousel, reel, story, idea, script or ad.
- `scheduled_at`: New date and time in ISO 8601, e.g. "2026-11-03T18:00:00Z".
- `status`: New status: draft, in_review, approved, scheduled or published. Never publishes anything.
- `title`: New title.

### `xcut_delete_calendar_post` (~70 tokens)

Delete a calendar post

Use this when the user explicitly asks to remove a post from their XcutAI content calendar. Do not use to reschedule or edit (use xcut_update_calendar_post).

Input parameters:

- `post_id` (string, required): Id (UUID) of the calendar post to remove, from xcut_list_calendar.

### `xcut_get_breakouts` (~96 tokens)

Get breakout alerts and trends

Use this when the user asks what is breaking out or trending in their own niche right now: XcutAI's breakout alerts from the accounts on their watchlist, plus trending keywords, from the last scan. Free. Do not use for a topic or creator outside their watchlist (use xcut_search_viral).

Input parameters:

- `limit` (integer): Most breakout alerts to return, 1-50. Default 20.

### `xcut_scan_watchlist` (~100 tokens)

Scan watched accounts now

Use this when the user wants fresh breakouts now instead of waiting for the daily scan. Rescans the accounts on their XcutAI watchlist; allowed once an hour; about 1 credit or less. Set `estimate` to true to get the cost without scanning. Do not use to read the latest results (use xcut_get_breakouts).

Input parameters:

- `estimate` (boolean): true returns the credit cost without scanning. Default false.

### `xcut_search_viral` (~248 tokens)

Search the viral library

Use this when the user asks what is going viral, trending or performing right now on Instagram, TikTok, YouTube or Facebook for a topic or niche, what a creator posts that works, or what ads run in the Meta Ad Library. `mode` is keyword (a topic) or profile (an account). Returns posts with views, engagement, links and captions, plus `next_cursor` for more. A fraction of a credit per page (cached pages free). Do not use for general web or news search, or for one profile's outliers versus its own normal (use xcut_find_outliers).

Input parameters:

- `cursor`: next_cursor from the previous page.
- `limit` (integer): Posts per page, 1-30. Default 15.
- `mode` (string): keyword to search a topic (e.g. "meal prep"), or profile to see an account's top posts (query = @handle). Default keyword. Facebook supports profile only.
- `platform` (string, required): instagram, tiktok, youtube, facebook, or meta_ads for the Meta Ad Library.
- `query` (string, required): A topic, or a handle when mode is profile.

### `xcut_find_outliers` (~217 tokens)

Find a profile's outlier posts

Use this when the user asks which of a creator's, competitor's or their own posts performed best or beat the account's normal, or what works for an account before writing. Scores the latest 20, 30, 50 or 100 posts against that account's own median and says by how much each beat it. About 1-2 credits for 30 posts. Longer runs return a `job_id` for xcut_get_job. Do not use for a full metrics table (use xcut_profile_analytics) or for a topic across many accounts (use xcut_search_viral).

Input parameters:

- `canvas_id`: Canvas (UUID) that owns the thumbnails. Optional. The results are returned here, not placed on the canvas.
- `platform` (string, required): instagram, tiktok, youtube or facebook.
- `posts` (integer): How many recent posts to analyse: 20, 30, 50 or 100. Default 30.
- `profile` (string, required): Profile link or @handle.

### `xcut_profile_analytics` (~179 tokens)

Get profile analytics

Use this when the user wants an account's per-post metrics (views, likes, comments, shares, engagement) over the last N posts or days, for a performance review or to compare accounts on Instagram, TikTok, YouTube or Facebook. Credit cost depends on platform and post count. Longer runs return a `job_id` for xcut_get_job. Do not use to find standout posts (use xcut_find_outliers).

Input parameters:

- `canvas_id`: Canvas (UUID) that owns the thumbnails. Optional. The table is returned here, not placed on the canvas.
- `days`: Or: posts from the last N days.
- `platform` (string, required): instagram, tiktok, youtube or facebook.
- `posts`: Last N posts (default 20).
- `profile` (string, required): Profile link or @handle.

### `xcut_get_job` (~94 tokens)

Check a job

Use this when another XcutAI tool returned a `job_id`: it returns status `running` with a suggested wait, or the finished result (outlier analysis, profile analytics, imports with their transcript, agent replies). Free. Do not use without a `job_id` from an XcutAI tool.

Input parameters:

- `job_id` (string, required): The `job_id` an XcutAI tool returned, unchanged.

### `xcut_import_post` (~197 tokens)

Import a social post

Use this when the user shares an Instagram, TikTok, YouTube or Facebook post or video link and wants its transcript, a breakdown of why it worked, or their own version of it. XcutAI transcribes the video and reads the caption, visuals and stats. Don't ask which canvas: with none named it goes to the user's "From AI apps" canvas. A fraction of a credit plus transcription (1-2 credits for a short video). Returns a `job_id`; xcut_get_job returns the transcript when done. Do not use for website links (use xcut_import_website) or Meta Ad Library links (use xcut_import_meta_ad).

Input parameters:

- `canvas_id`: Optional. Leave it out unless the user names a canvas: XcutAI then uses their "From AI apps" canvas. Ids come from xcut_list_canvases.
- `url` (string, required): Post or video link.

### `xcut_import_website` (~129 tokens)

Import a website

Use this when the user shares a web page (product page, landing page, article) to use as context for writing their content. XcutAI reads the page and stores its text on a canvas. Up to about 3 credits. Returns a short preview. Do not use for social media post links (use xcut_import_post).

Input parameters:

- `canvas_id`: Optional. Leave it out unless the user names a canvas: XcutAI then uses their "From AI apps" canvas. Ids come from xcut_list_canvases.
- `url` (string, required): Page link.

### `xcut_import_meta_ad` (~146 tokens)

Import a Meta ad

Use this when the user shares a Meta Ad Library link (facebook.com/ads/library) and wants the ad's copy, media and video transcript to study or remix. Holds up to 18 credits and charges the real cost. May return a `job_id` for xcut_get_job. Do not use for organic Facebook posts (use xcut_import_post).

Input parameters:

- `canvas_id`: Optional. Leave it out unless the user names a canvas: XcutAI then uses their "From AI apps" canvas. Ids come from xcut_list_canvases.
- `url` (string, required): Meta Ad Library link (facebook.com/ads/library/?id=…).

### `xcut_import_profile` (~174 tokens)

Import a profile's recent posts

Use this when the user wants to study a creator's or brand's recent posts in depth (Instagram, TikTok, YouTube or Facebook): it adds them to a canvas as a group and transcribes and analyses each. About 1-15 credits plus transcription. Returns a `job_id` while the profile is read. Do not use just to rank posts (use xcut_find_outliers).

Input parameters:

- `canvas_id`: Optional. Leave it out unless the user names a canvas: XcutAI then uses their "From AI apps" canvas. Ids come from xcut_list_canvases.
- `platform` (string, required): instagram, tiktok, youtube or facebook.
- `posts` (integer): How many recent posts (default 6).
- `profile` (string, required): Profile link or @handle.

### `xcut_add_note` (~135 tokens)

Add a note to a canvas

Use this when the user wants a script, hook list, plan or other text saved to XcutAI. Accepts Markdown; with no canvas named it goes to the user's "From AI apps" canvas. Free. Do not use for calendar posts (use xcut_add_to_calendar).

Input parameters:

- `canvas_id`: Optional. Leave it out unless the user names a canvas: XcutAI then uses their "From AI apps" canvas. Ids come from xcut_list_canvases.
- `text` (string, required): Markdown text.
- `title`: Optional note title, e.g. "Week 1 scripts".

### `xcut_get_content` (~158 tokens)

Read an imported item

Use this when you need the full content of an item on the user's XcutAI canvas, by canvas id and item id (from xcut_get_canvas): a post or video's transcript and stats, a note, a chat's messages (paginate with `before`), a website's text, an analytics table, a Meta ad, or a group's contents. Free. Do not use to list a canvas (use xcut_get_canvas).

Input parameters:

- `before`: Chats only: message id to page back from.
- `canvas_id` (string, required): Canvas id (a UUID) from xcut_list_canvases.
- `item_id` (string, required): Item id from xcut_get_canvas, or the `item_id` an import returned.

### `xcut_write_content` (~251 tokens)

Write with an XcutAI agent

Use this when the user wants XcutAI's agents to write for their brand: Reels, TikTok and Shorts scripts, YouTube scripts, Facebook posts, Meta ad copy, hooks, captions or plans, using their brand memory, imported posts and XcutAI's library of formats that performed. The conversation is saved to their canvas. Pick `agent` for the platform, or auto. Token-billed (typically 1-26 credits). Longer turns return a `job_id`. Do not use for writing unrelated to the user's social media or marketing content.

Input parameters:

- `agent` (string): Which XcutAI writer to use: auto (picks from the request), instagram, tiktok, youtube_script, youtube_shorts, facebook, meta_ad or general. Default auto.
- `canvas_id`: Optional. Leave it out unless the user names a canvas: XcutAI then uses their "From AI apps" canvas, where imports also land. Ids come from xcut_list_canvases.
- `item_ids`: Only use these canvas items (default: the whole canvas).
- `request` (string, required): What to write, as the user would ask in the canvas chat.

### `xcut_run_playbook` (~247 tokens)

Run an XcutAI playbook

Use this when the user wants a ready-made content deliverable from what is on their canvas: a month or week content plan, 25 hooks, 3 scripts, breakout remix, competitor teardown or benchmark, prospect audit, ad-angle gap map, ad creatives, carousel or thumbnail concepts. Token-billed (a month plan is roughly 60-90 credits). Longer runs return a `job_id`. Do not use for a single custom request (use xcut_write_content).

Input parameters:

- `canvas_id`: Optional. Leave it out unless the user names a canvas: XcutAI then uses their "From AI apps" canvas, where imports also land. Ids come from xcut_list_canvases.
- `days`: Plan length, for month/week plans.
- `notes`: Extra instructions from the user.
- `platform`: The platform to write for, when the playbook follows one.
- `playbook` (string, required): month or week (content plan), hooks (25 hooks), scripts (3 scripts), remix (breakout remix), teardown, benchmark, audit (prospect audit), angles (ad-angle gap map), adcreatives, carousel or thumbnail…

### `xcut_generate_image` (~187 tokens)

Generate an image

Use this when the user wants a social media image made or edited: post graphic, carousel slide, thumbnail or ad visual, from a prompt and optional reference image links. Needs the Starter plan or higher; about 14 credits per image. Do not use for images unrelated to the user's content or marketing.

Input parameters:

- `aspect_ratio` (string): 1:1, 4:5, 9:16, 16:9, 3:4 or 4:3. Default 1:1 (use 4:5 for Instagram feed, 9:16 for Reels and Stories).
- `prompt` (string, required): What to make or change, e.g. "bold carousel cover: '3 retinol myths', pink gradient, large serif headline".
- `reference_image_urls`: Up to 5 public https image links to guide style or to edit.

## Diagnostics

Captured diagnostic sections: TLS, DNSSEC, Authorisation, Transports. The full working is on the page: https://verifymcp.io/servers/ai-xcut-xcutai/mcp#diagnostics

## Score history

- 2026-10-18: 72
- 2026-10-17: 71
- 2026-10-16: 71
- 2026-10-15: 71
- 2026-10-14: 70
- 2026-10-13: 70
- 2026-10-12: 69
- 2026-10-10: 68

## Common questions

### What is the XCut AI – Viral Content Research MCP server?

XCut AI – Viral Content Research is an MCP server listed in the public MCP registry as ai.xcut/xcutai. Find viral posts, outliers, hooks and ads on Instagram, TikTok, YouTube and Facebook, then write. This page covers its hosted endpoint (https://mcp.xcut.ai/mcp).

### Is the XCut AI – Viral Content Research MCP server safe to use?

XCut AI – Viral Content Research scores 72 out of 100 on VerifyMCP. That is a record of what we were able to check automatically, not an endorsement. The category breakdown on this page shows every signal behind the number, including the ones we could not confirm.

### What tools does the XCut AI – Viral Content Research MCP server expose?

XCut AI – Viral Content Research exposes 39 tools: xcut_find_hooks, xcut_get_playbook, xcut_hook_ideas, xcut_write_caption, xcut_rewrite_bio, and 34 more. Their descriptions and schemas cost roughly 5,743 tokens of context every time the server is loaded.

### Does the XCut AI – Viral Content Research MCP server require authentication?

No. We connected to XCut AI – Viral Content Research without credentials and it answered, so anything it exposes is reachable by anyone who knows the address.

### Is the XCut AI – Viral Content Research MCP server still maintained?

XCut AI – Viral Content Research is still listed as active in the MCP registry. We last reached this channel on 18 October 2026. Those dates come from our own scans of the registry and the channel itself, not from anything the publisher announced.

## Links

- Remote endpoint: https://mcp.xcut.ai/mcp
- Website: https://xcut.ai/mcp
- Changelog RSS feed: https://verifymcp.io/servers/ai-xcut-xcutai/mcp.xml
- Changelog JSON feed: https://verifymcp.io/servers/ai-xcut-xcutai/mcp.json
- HTML version of this page: https://verifymcp.io/servers/ai-xcut-xcutai/mcp