Metricool API: Plans, Token, Endpoints, Limits and Integrations

Who gets the Metricool API, and what it costs

Metricool's public REST API is the programmatic door into the same account you use in the app: the planner, analytics, competitors, the inbox, reports, SmartLinks and campaign dashboards. It is available only on the Advanced and Custom plans. Free and Starter have no API access at all, and the help center tells those users to upgrade first ^[1]. The App Store plan terms do not list it either. One older marketing page places the API on Custom alone, but the pricing page and the newer help pages put it on Advanced too, and we follow them. We read the API guide, the help center and the Terms; we did not build an integration ourselves.

Advanced costs $67 a month billed monthly or $53 a month billed annually ($636 a year), for up to 15 brands. The same plan runs $107 or $85 a month for 25 brands, and $210 or $159 a month for 50 brands. Custom covers more than 50 brands and has no published price. Starter, at $25 a month billed monthly or $20 a month billed annually for up to 5 brands, does not unlock anything here. On plans bought on the web, X is not included: it needs a paid plan plus the X add-on, $10 a month per connected X account for buyers since July 2026 (the app-store Starter Mobile 5 plan includes X in its price). The full plan math sits in our Metricool pricing guide.

Finding your token, user ID and brand IDs

Every call needs three things: your access token, your userId and the blogId of the brand you are working on. The token lives in Account Settings, under API. That same screen shows your current usage, the history of endpoints you have called and a link to the documentation. The token is not the same thing as your numeric user ID, and the help center keeps the two apart on purpose. If you build a Looker Studio report, the identical token also appears under Reporting, in the Data Studio section, because the connector runs on it.

Brand IDs come from the API itself. The simpleProfiles endpoint returns every brand you manage, plus every brand shared with you, each with its blogId. A brand's ID stays the same even when you swap the profiles connected to it, so scripts keyed to a blogId survive a reconnection. Only the account owner can see the token. That matters more than it sounds. The token opens every brand linked to the account, so Metricool tells agencies never to hand it to a client. Shared brands cannot use the Looker Studio connector for the same reason: the token belongs to the owner alone.

Making your first call

The authentication model is simple. You send the token in a header named X-Mc-Auth, and you pass userId and blogId as parameters on the request. The base address for every endpoint is app.metricool.com/api, according to the basic API guide and the white-label technical pages. Two older help articles disagree. The n8n page shows a different host with a v1 path, and the Power BI page describes an Authorization header. We follow the API guide, which is more specific and matches the white-label technical pages, and so should you. If a request fails with an authentication error, check the header name before anything else.

Metricool's own API guide walks through the process in 8 chapters, from the token to a working integration, and it starts with a first call that changes nothing, so you can test credentials safely. The guide also notes the API works from code, from automation tools like Make (the automation tool) or n8n, and from AI coding assistants like Claude Code. A sensible order for a first session looks like this:

1. Confirm the account is on Advanced or Custom, then copy the token from Account Settings, API.
2. Note your userId and call simpleProfiles to list brands and their blogIds.
3. Send a read-only request with the X-Mc-Auth header and both IDs.
4. Check the usage screen to see the call registered, then move on to writes.

The technical reference lives at app.metricool.com/resources/apidocs/index.html, downloadable as swagger.json or swagger.yaml, and there is an English PDF too. Some endpoints are missing from the PDF. Metricool suggests finding those by inspecting the browser while you use the app. That works, but undocumented routes can change without notice, so keep anything built on them behind a thin wrapper in your own code. Prefer the swagger file as your source of truth, because it can feed a client generator or a Postman collection directly.

What the endpoints cover

Scheduling posts

The route groups cover /v2/scheduler, /v2/analytics, /stats, /v2/analytics/competitors, /v2/inbox, /v2/reporting, /v2/smart-links and /v2/reporting/campaigns-dashboards. The scheduler is the deepest. It supports create, list, get, update, patch and delete for scheduled posts, plus bulk send-to-review and approve or reject. A new post goes to POST /v2/scheduler/posts with publicationDate, text and providers as the minimum fields, and the date must be in the future. Optional fields include autoPublish, draft, creatorUserMail, media and network-specific data. Media must come from public, non-expiring URLs; private or temporary links are skipped silently and the post is scheduled without them. Posts with media also need saveExternalMediaFiles set.

Autolists, usage and media quirks

Adding content to an autolist takes two calls: one creates an empty slot, the second fills it with text and media. The create-slot endpoint only accepts GET, and a POST returns a 500 error. The fill call goes as form-data, with pictures set to an empty array when there is no image. Watch the first step. Every create-slot call adds a new empty row, even if you never finish the second one. Two media details round this out: images imported through the API are not converted to JPG, and iPhone HDR video is not supported when Metricool publishes by API, so it can come out yellowish. Usage statistics can be pulled through their own endpoint, which is handy for monitoring.

The limits that shape your design

No published call limit on Advanced

We found no published call limit for the API on Advanced. Under a Custom-plan Integration or White Label agreement, Metricool's Terms define standard API use as up to 500 calls a month per brand, and above that it may raise the agreed rates after a quote ^[2]. On one of those contracts that is a modest budget: a script that polls analytics every hour for one brand would blow through it within a month. Either way, design for batch pulls, cache what you read and keep an eye on the usage screen. The Terms also prohibit scraping Metricool data without written authorization, so the API is the sanctioned route.

No calls from the browser

The API is not built for front-end code. Browsers that call it directly fail, because the API answers the CORS preflight with a 403 ^[3]. Your own backend has to receive the browser's request, attach the credentials and forward it. That rule is also a security one. If the token ever leaks into front-end code, Metricool's advice is to regenerate it from Account Settings right away.

Inbox comments: only the 50 most recent posts

Comments can be exported through /v2/inbox/post-comments, one call per network, for Facebook, Instagram, LinkedIn, YouTube and in principle TikTok. The API returns comments for the 50 most recent posts per network (on YouTube, the 50 most recent comment threads), and the limit counts posts, not dates ^[4]. Comments are not available for X, Threads, Pinterest or Bluesky. Older posts fall out of reach.

What no API can do: the networks' own rules

Many gaps are not Metricool's choice. Each network decides what third-party tools may do through its API, and Metricool cannot offer what the network does not expose. Published posts cannot be edited from Metricool, on Instagram, LinkedIn, Threads or Bluesky alike. Publishing straight into Facebook Groups stopped when Meta withdrew the permission on 22 April 2024; only sharing a link to a group still works. YouTube community posts cannot be published because the YouTube Data API has no public endpoint for them.

The smaller limits add up. Instagram allows no GIFs, no Story add-ons, no music in carousels or feed posts, and it publishes every video as a Reel. LinkedIn rules out articles, event posts and co-authored posts. TikTok takes no line breaks, overlay text or voiceovers, and Facebook caps API mentions at 3 pages. Volume caps run over a rolling 24 hours: 50 posts on Instagram, 30 Reels on Facebook and 250 posts and 1000 replies on Threads. YouTube adds a cap that is Metricool's own, not the network's: 10 posts on Free or 20 on paid plans. Build these limits into any bulk scheduler, or expect failed posts.

No-code routes: Zapier, Make, n8n and Power BI

Zapier

The Zapier integration runs on the same token and is limited to Advanced and Custom. It has six triggers: New Failed Post, New Draft in Planner, New Scheduled Post, New conversation in Inbox, New Published Post and New GMB review. It has four actions: Add a post to an Autolist, Schedule a post, Reply to a conversation and Reply to a GMB Review. Multi-post threads cannot be scheduled through Zapier. The inbox trigger covers messages and comments on Instagram, Facebook and X profiles.

Make, n8n, Power BI and Tableau

Make (the automation tool) connects with the API key pasted from account settings, so it also needs Advanced or Custom. Metricool publishes four English-only video tutorials for it, covering metrics, scheduling, Google Sheets and direct API calls. There is no official n8n integration: n8n users call the API with the HTTP Request node, and Metricool warns this takes some technical skill. Power BI can pull social stats, ad campaign metrics for Facebook, Instagram, Google and TikTok Ads, and inbox data through the API. Tableau works too, through a Web Data Connector. In practice, any tool that sends HTTP requests with custom headers, Postman included, can talk to the API.

Looker Studio: the reporting route without code

If all you want is dashboards, the Looker Studio connector (Metricool still calls it Data Studio) may save you from writing code at all. It needs an Advanced or Custom account and uses the API token from account settings. It covers all your brands and can blend them statically by selection or dynamically by pattern. Metricool sets no limits on responsible use of the connector beyond the brands in your plan. Google allows only one access credential per account for this connector, so several API keys need several Google accounts.

Sharing is where people trip. A client viewing a report needs the data source set to owner credentials; viewer credentials would require your exact token, which you should never share. Share the report, not the key. The connector has kept growing since Google approved it for the official community list in 2020, and Threads and Bluesky data were added later. Expect the E012 error after you regenerate a token, switch accounts or change plans, since the old token is no longer valid.

White label for integrators: the Custom-only endpoints

White Label for Integrators is a Custom feature, and it adds endpoints of its own. The /admin/ routes add, delete, edit and consult brands for your end clients. GET /admin/add-profile creates a brand from your web app without the end user ever visiting Metricool. Only the master account can create brands this way, and an empty response means there is no room for another brand or a collaborator tried. Brand creation is a GET with a side effect, so a browser prefetch, a crawler or a double click can create duplicates; add a confirmation step.

Embedded sessions run on autoin tokens from /api/v2/settings/whitelabel/logininfo/{blogId}, with a READONLY or MANAGER scope. They last 8 hours (480 minutes) by default and can be set as short as 30 minutes. They cannot be reused, so generate a new one each time. Use the brand's White Label token in the autoin link, not the general token, or the iframe will not display. An anonymous-link endpoint lets end clients connect their own accounts without a Metricool login; only one link per brand is valid at a time. That endpoint's own help page lists Advanced or higher, while the integrator pages call White Label for Integrators Custom only. End clients get no Zapier, Make or Looker Studio, and there is no user management, so the integrator builds its own sign-in.

Regenerating the token without breaking everything

Since June 2026, Advanced and Custom accounts can regenerate the API token from Account Settings. The old token dies immediately. Every integration using it stops: Looker Studio, Zapier, Make, n8n and your own scripts, until you paste in the new one. That is the right move after a leak, but it needs planning. Before you click, list every place the token lives, then update them in one sitting. Agencies with many reports should expect a burst of E012 errors in Looker Studio until each data source is reconnected.

API or MCP: two different products

Metricool also runs an official MCP server, and the two are easy to confuse. The MCP works on every plan, including Free, while the API stays on Advanced and Custom. The MCP lets an AI client such as Claude or ChatGPT read and schedule through a sign-in, in plain language; plan limits still apply to whatever it does. The API is for programmatic export and automation from your own code or workflow tools, and Metricool recommends each for exactly those jobs. The API also reaches data the MCP does not: inbox comments are available only through the API. Our Metricool MCP guide covers the setup.

One wrinkle connects them. n8n reaches the MCP with an X-Mc-Auth header carrying the API key, and any token-based MCP connection that uses the API requires Advanced. So an n8n workflow does not escape the plan gate, even though the MCP itself is open to everyone. Rivals draw the line elsewhere. Buffer does not gate its API behind a paid tier: even its free tier gets 1 key and 3,000 requests per 30 days, capped at 250 per 24 hours. Later has no public API at all.

The bottom line

Metricool's API is a capable, conventional REST interface with one header, two IDs and a well-documented scheduler, but it is gated. You need Advanced or Custom, every call has to go through a backend, and comment exports stop at the 50 most recent posts per network. If you only want reports, the Looker Studio connector does the job without code. If you want an AI assistant on your data, the MCP is cheaper and open on every plan. Build against the API when you need real automation, and design around its limits from day one.