Which trend research tools can Claude or ChatGPT use through MCP?
Short answer
On 8 November 2026 seven trend research tools published an official MCP server that Claude, ChatGPT and other clients can connect to: vidIQ, Eden, Sandcastles, 1of10, ViralMaxing, Virlo and OutlierKit. Only vidIQ includes its MCP on a free plan (150 credits a month). They differ in which networks they search and in what a video is measured against, so the same video can be a big outlier in one tool and an ordinary one in another. None of them publishes posts; that is a separate tool.
The seven, side by side
Every cell is what the vendor’s own page said on 8 November 2026, signed out. Prices are the monthly list price in the vendor’s currency; vendors change them, and some show a local price at checkout.
| Tool | Networks it searches | A video is measured against | MCP server | Cheapest way to use the MCP |
|---|---|---|---|---|
| vidIQ | YouTube (long videos and Shorts), Instagram Reels, and TikTok for outlier research (help) | The channel’s usual baseline, with views and velocity (Outliers) | mcp.vidiq.com/mcp | Free plan: 150 credits a month; most research tools cost 5 credits a call (plans) |
| Eden | YouTube, TikTok, X, Instagram, LinkedIn, Substack | The creator’s own median | mcp.eden.so/mcp | Connecting is free, but the Free plan has no AI credits; credits start on Personal Plus, $15 (pricing) |
| Sandcastles | TikTok, Instagram, YouTube Shorts; no long YouTube videos (help) | The median views of the channel’s videos from the 90 days before (glossary) | mcp.sandcastles.ai | Pro, $49 (plans) |
| 1of10 | YouTube long videos and Shorts | The channel’s average | Shown inside the account | Pro, $89 (pricing) |
| ViralMaxing | TikTok, Instagram Reels, YouTube Shorts (search) | The creator’s usual result, plus a 0–100 VM Score whose formula is not published | api.viralmaxing.com/api/mcp | No free trial; Standard from 2,680 ₽, half that for the first month (pricing) |
| Virlo | TikTok, YouTube Shorts, Instagram Reels, Meta ads (its FAQ still lists Instagram Reels as planned) | The creator’s follower count (Outliers) | dev.virlo.ai/api/mcp/mcp | Pay as you go, from a $10 balance |
| OutlierKit | YouTube | The channel’s historical baseline: average, median and spread (finder) | outlierkit.com/mcp | Its pages disagree: any paid plan in the docs, Pro or higher on the MCP page; Hobby is $49 and Pro $99 (pricing) |
Two filters people ask for are uneven. A language filter is documented by Eden (in its API) and by 1of10; vidIQ, Sandcastles and OutlierKit list their filters without one. A country filter is documented only by Virlo, for its topic trends.
Why the same video gets different scores
An outlier is a video that did far better than its creator usually does, and the tools disagree on what “usually” means. The arithmetic and its trade-offs are in What is an outlier video?; here is what it does to a comparison.
- Median against average. Eden and Sandcastles divide by a median. 1of10 says it measures against the channel’s average, and ViralMaxing’s documentation says the same of its multiplier. One earlier hit pulls an average up, so the same video scores lower against an average than against a median.
- The window. Sandcastles counts only the channel’s videos from the 90 days before the one being scored, so the baseline moves with the channel. A tool with a longer or undated baseline gives an older channel’s new video a different score.
- Views against followers. Virlo compares views with the creator’s follower count, not with past views. A large account whose videos usually reach a fraction of its followers can look ordinary there and still be an outlier in a median-based tool.
- What counts as a video. Sandcastles leaves out long YouTube videos; 1of10 and OutlierKit are YouTube only. A “top outlier this week” list from one tool is a list from its own index, not from the whole network.
What a free source gives an agent instead
An agent can do part of this without any of the seven, from the networks’ own sources, within their limits:
- YouTube. The Data API’s search has its own daily allowance of 100 calls, and view counts come from separate calls against a 10,000-unit daily quota (quota calculator). YouTube’s policies restrict derived metrics such as a views-to-median ratio; the details are in the YouTube Shorts trends answer.
- TikTok. The Creative Center ranks hashtags by country for the last 7, 30 or 90 days; signed out it shows the top three (Trends). More in How do you find trends on TikTok?.
- Instagram. Meta’s official hashtag search allows 30 different hashtags in a rolling 7 days, needs App Review, and its recent list covers 24 hours (Hashtag Search). There is no official list of other people’s popular reels; see How do you find trending reels on Instagram?.
What their terms say about reusing results
If you build an agent that passes results on to other people, read the vendor’s terms first. vidIQ forbids reselling, sublicensing or commercially redistributing its service or materials (terms); Eden says plainly not to scrape or resell it (terms); Virlo forbids scraping, reselling or redistributing its data outputs outside a paid API licence (terms); ViralMaxing forbids mass scraping through its API and building competing services (terms).
Where Mellow Hub fits
Since 8 November 2026, Mellow Hub has its own two trend tools on the same MCP server it publishes through: trends_search (a topic on YouTube, TikTok or Instagram, over a period you choose) and trends_breakdown (why one video worked, and your version with a caption). A YouTube search is free and shows YouTube’s own figures; a new TikTok or Instagram search costs 2 AI credits and carries ×N. They are not in the table above because the seven were compared on 8 November, before Hub’s shipped. Hub is also the step after: once you have made your own post from what you learned, Hub checks it against each network’s rules and publishes or schedules it to Instagram, TikTok, YouTube and the other networks you connect, through its own MCP server or REST API. vidIQ notes that its directory and key connections are read-only, which is typical: the research tools read, a publishing tool writes.
The two are separate connectors in the same client: one to research, one to publish. In Claude, check your plan first: Free allows one custom connector. The setup for Hub is in the social media MCP guide.
Hub has no free plan: a new account starts with a 7-day free trial of a monthly plan. Starting it takes a card, and $0 is charged that day; cancel before the 7 days end and nothing is charged. After the trial the plan renews monthly, from $19 a month (Solo), and the first paid month is 20% off. A yearly plan is charged on the day you choose it, with no free trial. Publishing is unlimited on every plan, under a fair-use ceiling that stops automated floods (900 publications a month on Solo), where one publication is one post on one account, so a post sent to five networks counts as five toward it. There are no automatic overage charges.
Questions people also ask
Is there a free trend research MCP?
vidIQ’s MCP is included on its Free plan, which gives 150 credits a month, and vidIQ says most research tools cost 5 credits a call. Eden’s MCP is free to connect, but its Free plan has no AI credits, so research calls start on a paid plan. The others need a paid plan or a balance.
Which of these tools covers long YouTube videos?
vidIQ, 1of10 and OutlierKit cover long YouTube videos, and Eden’s API has separate types for Shorts and other YouTube videos. Sandcastles says it does not import long-form YouTube videos, only Shorts. ViralMaxing and Virlo name YouTube Shorts.
Can one of these tools post the video for me?
Not in what their pages describe: they read and analyse. vidIQ says its directory and API-key connections are read-only. Publishing is a separate tool, such as a scheduler with its own MCP server, connected next to the research one.
Why does each tool give the same video a different outlier score?
They divide by different things: the creator’s median, the channel’s average, the median of the last 90 days, or the follower count. Each tool also scores only the videos in its own index. Compare scores within one tool, not across tools.
Do I need an MCP tool at all to find outliers?
No. You can compare a video’s views with the creator’s usual views by hand, and an agent can do it for YouTube with the official Data API within its quota and policies. The paid tools add an index across many creators, and, for TikTok and Instagram, data the official APIs do not give.
Sources
- vidIQ: MCP server
- vidIQ Help: vidIQ MCP
- vidIQ: Plans
- vidIQ: Outliers
- Eden: Discover
- Eden: MCP
- Eden: Pricing
- Sandcastles Help: Glossary
- Sandcastles Help: Videos filter settings
- Sandcastles: MCP
- Sandcastles Help: Plans
- 1of10: MCP
- 1of10: Pricing
- 1of10: Features
- ViralMaxing: MCP
- ViralMaxing: Pricing (Russian edition)
- ViralMaxing: Viral Reels search
- ViralMaxing: Documentation for AI assistants (llms-full.txt)
- Virlo: Outliers
- Virlo: MCP
- Virlo: FAQ
- OutlierKit: MCP docs
- OutlierKit: MCP server
- OutlierKit: Outlier finder
- OutlierKit: Pricing
- vidIQ: Terms of Service
- Eden: Terms
- Virlo: Terms of Service
- ViralMaxing: Terms
- YouTube Data API: Quota calculator
- TikTok Creative Center: Trends
- Meta for Developers: Instagram Hashtag Search
Other companies’ features change. Every statement about them rests on these pages, as they read on the date above.