Menu
Mattering Automations connector

A Moz MCP that tells you what each call costs before you run it

Connect one Moz API token and 31 nodes land on the canvas: Domain Authority and Brand Authority, backlinks and linking domains, anchor texts, top pages, ranking keywords and the full keyword research set. The Moz API is metered in rows, so every node here states its row cost in its own description, four nodes are free, and the connection test spends nothing.

31 ready-made nodes One credential to connect Works with Claude, ChatGPT, and Matter

What the Moz MCP connector does

The Moz connector for Mattering Automations is 31 nodes on Moz's current unified API, the JSON-RPC surface at api.moz.com/jsonrpc, authenticated with one token sent as the x-moz-token header. Site metrics carry Domain Authority, Page Authority, Spam Score and the full link-count matrix. The link nodes carry backlinks, anchor texts, linking domains, recently gained and recently lost domains, link status checks and link intersects. Keyword research carries volume, difficulty, organic CTR, priority, search intent, related keywords and ranking keywords. Thirty nodes read and one, Moz: Custom API Method, is the escape hatch to any Moz method this catalog does not ship a typed node for.

Money is the first thing to understand about this API, so it is the first thing this page says. Moz meters in ROWS. In Moz's own words, one row typically corresponds to one object in the response for methods that return multiple objects, or one field in the response for methods that return a single flat object or report. The practical translation is short: on a list node, the page size IS the price. Moz: List Links at a page size of 25 costs 25 rows. Moz: List Ranking Keywords allows a page size of 500, which is 500 rows in one click. Moz: Fetch Link Intersects is the one that surprises people, because it charges rows equal to the number of targets for every result returned, so 25 results across 6 targets is 150 rows. Each node repeats its own cost shape in its description, and there are only four shapes: one row per call, one row per returned result, one row per returned value up to four on Moz: Fetch Keyword Metrics, and the per-target multiplier on link intersects.

Four nodes are free of charge and Moz says so itself: Moz: Fetch Quota, Moz: Fetch Current Index ID, Moz: Get Top Ranking Domains and Moz: Get Top Ranking Pages. Those four are what make a metered connector safe to automate. Put Moz: Fetch Quota in front of an expensive branch and gate on used against allotted, so a scheduled workflow stops itself instead of running into a limit. Poll Moz: Fetch Current Index ID, which changes only when Moz refreshes its link index, and run the expensive link report only when it moves. Neither of those costs a row, and both are workflows rather than questions, which is the part a chat window cannot hold for you.

Every node speaks MCP, the open protocol for connecting AI assistants to tools, so Claude, ChatGPT, and any MCP-capable assistant can use them, alongside Matter, the AI assistant built into the platform. An agent can read Moz: Fetch Quota before it does anything else, pull Moz: Fetch Site Metrics for a prospect, ask Moz: List Recently Lost Linking Domains what disappeared this month, and write the answer into a task or a database, without anyone opening the Moz dashboard. The same nodes run unattended on a schedule, which is the difference that matters on a metered API: the guardrail runs whether or not a person is watching.

Moz ships its own MCP servers and this page is not going to pretend otherwise. Moz documents two endpoints, one for Moz Data and one for Moz Local, each requiring a valid subscription to the corresponding product, and both are published in Claude's connector directory. Moz describes the server as being in beta with access limited to account owners today, and says it is working to expand access to seated users. If what you want is to chat with your Moz data, use it. This connector is a different job: 31 typed nodes that run on a schedule or a trigger with nobody present, chain with every other connector in one workflow, hold the token once in a shared workspace store rather than in one person's client config, and state their row cost on the node so a workflow can be reviewed before it spends anything.

One connection, three surfaces

  • Visual automations. Drag Moz nodes onto the canvas and wire them to triggers, conditions, and other apps.
  • AI agents, yours or ours. These nodes speak MCP, the open protocol, so Claude, ChatGPT, and any MCP-capable assistant can use them, and Matter, the agent built into the platform, uses them natively.
  • The whole platform. Combine with native tasks, projects, and other connectors, no middleware in between.

Every Moz node, ready to drop on the canvas

31 nodes, generated from the connector itself so this list always matches what ships. Expand a category to browse.

Read and query nodes 30

Moz: Count Ranking Keywords

How many top-50 keywords a target ranks for, as a single total plus a breakdown by each rank from 1 to 50. The cheap way to size a competitor's organic footprint before deciding whether to spend rows listing it. SPENDS QUOTA: uses 1 row per call, whatever the totals turn out to be.

Moz: Fetch Brand Authority

Brand Authority score for a domain: Moz's measure of how strong a brand's presence is in search, as distinct from how many links point at it. Useful for competitor benchmarking and for tracking PR impact. SPENDS QUOTA: uses one row per call.

Moz: Fetch Current Index ID

An opaque id that changes whenever Moz's link index is refreshed. Store it and compare: when it moves, the link data underneath every other node in this connector has changed, which is the cue to re-run a link-building report rather than polling one on a fixed schedule. FREE: does not require any quota.

Moz: Fetch Keyword Difficulty

Keyword Difficulty for one keyword: how hard the top of that SERP is to break into. Use Fetch Keyword Metrics instead if you also want volume or CTR. SPENDS QUOTA: uses one row per call, so a loop over a keyword list costs one row per keyword.

Moz: Fetch Keyword Metrics

All four keyword metrics at once for one keyword: monthly volume, difficulty, organic CTR and priority. Cheaper than four separate calls when you want more than one of them. SPENDS QUOTA PER VALUE: one row per returned value up to a maximum of four, with nulls not charged, so a keyword Moz has full data for costs 4 rows.

Moz: Fetch Keyword Organic CTR

Organic CTR (Moz calls this Opportunity) for one keyword: how much of that SERP's traffic the organic results actually get, once ads and SERP features have taken their share. SPENDS QUOTA: uses one row per call.

Moz: Fetch Keyword Priority

Priority for one keyword: Moz's blend of volume, difficulty and organic CTR into a single go-after-this score. SPENDS QUOTA: uses one row per call.

Moz: Fetch Keyword Search Intent

Why someone searches a keyword: the primary intent labels plus a scored breakdown across informational, navigational, commercial and transactional. Feeds content briefs and page type decisions. Accepts a narrower set of locales, devices and engines than the other keyword methods. SPENDS QUOTA: uses one row per successful call.

Moz: Fetch Keyword Volume

Monthly search volume for one keyword. SPENDS QUOTA: uses one row per call, so a loop over a keyword list costs one row per keyword.

Moz: Fetch Link Intersects

Link intersect: sources linking to some targets and not others, the Venn-diagram query behind competitor link-gap analysis. Each row names the source page or domain, its Domain Authority and Spam Score, and matching_target_indexes pointing back into the Is Linking To array. THE MOST EXPENSIVE NODE IN THIS CONNECTOR: it uses N rows of quota per result, where N is the total number of targets across Is Linking To and Not Linking To, so 25 results against 6 targets costs 150 rows.

Moz: Fetch Multiple Site Metrics

The same metrics as Fetch Site Metrics for a batch of targets in one call. Results come back under results_by_site with per-target failures isolated in errors_by_site, so one bad URL does not fail the batch. SPENDS QUOTA PER ROW: uses one row per returned result, so the cost is the size of the array.

Moz: Fetch Quota

How much row quota the account has, has used, and when the period resets, plus whether overages are enabled. Put this in front of an expensive branch and gate on data.result.quota.used against data.result.quota.allotted. FREE: does not require any quota, which is what makes it the right auth probe as well.

Moz: Fetch Site Metric Histories

Historical series for chosen metrics on one target, as daily or monthly absolute values or as period-over-period deltas. Each requested metric comes back as a metric_name plus a metric_values array of {date, value} points. At least one of the four metric-list fields must be filled. There is no date range to ask for: Moz returns the history it holds, so this node cannot be windowed (see the incremental note at the foot of moz_endpoints.py). SPENDS QUOTA PER METRIC: one row for every metric named across the four lists.

Moz: Fetch Site Metrics

Domain Authority, Page Authority, Brand-safe Spam Score, crawl status and the full link-count matrix (pages and root domains linking to the page, subdomain and root domain, split by followed, nofollow, redirect and deleted) for ONE target. The workhorse node. SPENDS QUOTA: uses one row per call.

Moz: Fetch Site Metrics Distributions

Link-profile histograms for a target: how the links pointing at it are spread across the source's linking root domain count, Domain Authority and Spam Score. Every distribution is a ten-element array of counts, ordered lowest bucket first, named METRIC_BY_MEASURE. Sources with no Spam Score are excluded from the spam buckets. SPENDS QUOTA: uses one row per successful call.

Moz: Filter Links by Anchor Text

Backlinks to a target whose anchor text matches one exact string, or whose anchor text is empty when the field is left blank. Use it to audit over-optimized anchors or to find the links behind a brand mention. SPENDS QUOTA PER ROW: uses one row per returned result.

Moz: Filter Links by Domain

Every link from ONE source root domain to a target, sorted by descending source page authority (this method allows no other sort). The node for answering does site A still link to us, and from which pages. SPENDS QUOTA PER ROW: uses one row per returned result.

Moz: Get Final Redirect Target

Where a URL finally lands after Moz follows the redirects it knows about. The audit node for a domain migration or a bulk page move: it answers whether the redirect Moz's index believes in is the one you built. NOT a live fetch: it reports the index's view, which lags the web. SPENDS QUOTA: uses one row per call.

Moz: Get Link Status

Does a link from this source to this target exist in Moz's index? Returns a single boolean alongside both parsed queries. The node for verifying that a placed link is still live. SPENDS QUOTA: uses one row per call.

Moz: Get Multiple Link Status

Batch link-existence check: many source-to-target pairs in one call, answered under results_by_query with failures isolated in errors_by_query. The node a link-building campaign monitor loops over. SPENDS QUOTA PER ROW: uses one row per returned result, so the cost is the size of the array.

Moz: Get Top Ranking Domains

The highest-authority root domains on the web, globally, each with the full site-metrics object. Reference data for calibrating what a Domain Authority number means. FREE: does not require any quota.

Moz: Get Top Ranking Pages

The highest-authority individual pages on the web, globally, each with the full site-metrics object. Reference data for calibrating Page Authority. FREE: does not require any quota.

Moz: List Anchor Texts

The anchor texts people use when linking to a target, each with the number of external root domains and external pages using it. Good for brand-mention clustering and for spotting anchor-text over-optimization. SPENDS QUOTA PER ROW: uses one row per returned result.

Moz: List Linking Domains

Root domains linking to a target, one row per source domain rather than per link, each with the source's full site metrics plus how many of its pages target you and how many of those are nofollow, redirect or deleted. The node for referring-domain counts and outreach prospecting. SPENDS QUOTA PER ROW: uses one row per returned result.

Moz: List Links

Backlinks pointing at a target, one row per linking page, each carrying the full site-metrics object for BOTH the source and the target plus the anchor text, first and last seen dates, and the nofollow / redirect / rel-canonical flags. The core link-research node. SPENDS QUOTA PER ROW: uses one row per returned result, so Limit is the price of the call.

Moz: List Ranking Keywords

Keywords a domain, subdomain, subfolder or page ranks for in Google's top 50, with the ranking URL, the position, the keyword difficulty and the monthly volume. The competitor research node. SPENDS QUOTA PER ROW: uses one row per returned result, and this method allows a page size of up to 500.

Moz: List Recently Gained Linking Domains

Root domains that started linking to a target in the last 60 days, each with the day it happened (date_gained) and the source's full site metrics. Optionally narrow to a date range inside that window. The node behind a link-acquisition report. SPENDS QUOTA PER ROW: uses one row per returned result.

Moz: List Recently Lost Linking Domains

Root domains that stopped linking to a target in the last 60 days, each with the day the link disappeared (date_lost) and the source's full site metrics. The node behind a link-decay alert. SPENDS QUOTA PER ROW: uses one row per returned result.

Moz: List Related Keywords

Related keyword suggestions for a seed keyword, each with a relevance score. The ideation node for content briefs and cluster building. SPENDS QUOTA PER ROW: uses one row per returned result, so asking for 50 suggestions costs 50 rows.

Moz: List Top Pages

The strongest pages on a domain or subdomain, ranked by Page Authority, linking root domains or external linking pages, each row carrying the full site-metrics object. Filter to a 4xx status to find dead pages that still hold authority. SPENDS QUOTA PER ROW: uses one row per returned result.

Action nodes 1

Moz: Custom API Method

Call ANY Moz JSON-RPC method by name, for the methods this connector does not ship a typed node for: the whole Moz Local surface (locations, listings, networks, Google Business Profile insights, local grid rankings, accounts and groups) and any method Moz adds after this catalog was built. Moz Local methods need a Moz Local subscription as well as a Data API one. SPENDS QUOTA UNPREDICTABLY: the cost is whatever the named method costs, and some Local methods WRITE. local.location.cancel and local.location.close end a paid listing. Read the method's own documentation before running it.

Connecting Moz takes one credential

Moz authenticates every request with a single API token created in the Moz API dashboard, under Add Token. An account can hold up to five active tokens, and deleting one immediately breaks anything still using it, so create a token for automation rather than reusing a developer's. The Moz API is a separate subscription from Moz Pro and Moz Local, which is the most common source of confusion: a Moz Pro seat does not carry API access. Moz does publish a free tier for testing, 50 rows per month with a valid credit card required to open it, and Moz's help documentation adds a throttle of one request every ten seconds on that tier. Treat it as a way to prove a credential rather than a way to run a report, because a single call to Moz: List Links at the default page size spends half of the monthly allowance. Two plan boundaries are worth checking before you design around the whole catalog. Moz Local methods, reachable only through Moz: Custom API Method, need a Moz Local subscription in addition to the Data API one. And Moz documents its beta methods as drawing on a separate allowance available on the Starter Medium plan and higher, which does not affect this connector today, because no node here is a beta method.

1

Sign in to Moz and open the Moz API dashboard. Click Add Token, name it for the automation that will use it, and copy the value.

2

In Mattering Automations, open Connectors, click Add Connector, choose Moz, and create a credential with that token.

3

Click Test Connection. Mattering calls Moz's quota lookup method, which Moz documents as requiring no quota, so testing a credential never costs you a row. A pass proves the token, proves the account has a Data API subscription, and reports the remaining row allowance in the success message.

4

All 31 Moz nodes appear in the canvas palette. Run Moz: Fetch Quota first, for free, and read the used and allotted numbers before you build anything that loops.

Common uses for the Moz MCP

Gate every expensive report on the quota that is left

This is the workflow that pays for the connector. Moz: Fetch Quota is free and returns how many rows the account has, how many it has used, when the period resets and whether overages are enabled. Put it first in the workflow, feed used and allotted into a condition, and let the branch decide: enough headroom and the report runs, not enough and the run stops and opens a task instead. A metered API punishes silent loops, and the honest fix is a step that costs nothing and refuses to proceed. It is also the reason the connection test is free rather than merely cheap: the same method backs both.

Run link reports only when the link index actually moves

Moz: Fetch Current Index ID returns an opaque id that changes when Moz refreshes its link index, and it costs nothing to call. Store it, compare it on a schedule, and only when it changes do you spend rows on Moz: List Links, Moz: List Linking Domains or Moz: List Anchor Texts. A daily backlink report on a weekly index is the classic way to burn a monthly allowance on identical data. This pattern is a workflow rather than a question, which is exactly the shape that belongs in an automation rather than a chat prompt.

Weekly authority tracking for a client list

Moz: Fetch Multiple Site Metrics takes a batch of targets in one call and returns Domain Authority, Page Authority, Spam Score and link counts for each, with per-target failures isolated so one bad URL does not fail the batch. Moz: Fetch Brand Authority adds Moz's measure of brand presence in search, which moves for different reasons than link counts do. Schedule it, write the results into a Mattering database, and chart the trend beside your other client reporting. Price it before you schedule it: this node costs one row per returned result, so the size of your client list is the weekly bill.

Alert on links you lost, and report the links you won

Moz: List Recently Lost Linking Domains returns root domains that stopped linking to a target in the last 60 days, each with the day the link disappeared and the source's full site metrics. Moz: List Recently Gained Linking Domains is the mirror image, and it is the raw material for a link-acquisition report that writes itself. Both nodes cost one row per returned result, so a sensible limit plus a sort is a report and an unbounded pull is an invoice. Route a lost link from a high-authority domain into a task with an owner, and let the low-value ones accumulate in a database.

Competitor link-gap analysis, priced honestly

Moz: Fetch Link Intersects answers the Venn-diagram question behind outreach: which sources link to some of these targets and not others. Each row names the source, its Domain Authority and its Spam Score, and points back into the targets it matched. It is also the most expensive node in the connector, because it charges rows equal to the number of targets for every result returned, so a six-target query returning 25 results is 150 rows rather than 25. Use it deliberately, with a small result limit and a target list you actually chose, and put Moz: Fetch Quota in front of it.

Keyword briefs that carry intent, not just volume

Moz: Fetch Keyword Metrics returns volume, difficulty, organic CTR and priority for a keyword in a single call, which is cheaper than asking four separate nodes because it bills one row per returned value up to four, with nulls not charged. Moz: Fetch Keyword Search Intent adds the scored breakdown across informational, navigational, commercial and transactional, and Moz: List Related Keywords expands a seed into a cluster at one row per suggestion. Chain those into a content brief, drop it into a Productivity task, and the writer gets the intent split and the difficulty in the same card as the assignment.

Example: A weekly backlink report that refuses to overspend

Built on the Automations canvas in a few minutes, or described in plain language to Matter and built for you.

1

Trigger: a weekly schedule fires.

2

Moz: Fetch Quota runs first and costs nothing. A condition compares used against allotted and stops the run, opening a task, if the remaining rows are below what the rest of this workflow will spend.

3

Moz: Fetch Current Index ID runs, also free, and its value is compared against the id stored from last week. If the index has not refreshed, the run ends here and no rows are spent on data that has not changed.

4

Moz: Fetch Site Metrics reads Domain Authority, Page Authority and Spam Score for the target at a cost of one row.

5

Moz: List Recently Gained Linking Domains and Moz: List Recently Lost Linking Domains run with an explicit limit, because each costs one row per returned result and the limit is therefore the price of the step.

6

The rows land in a Mattering database next to last week's, a task opens only when a lost domain is above the authority threshold you set, and the new index id is stored for next week's comparison.

Moz MCP questions, answered

What does the Moz API cost to run through these nodes?

Moz meters in rows against a monthly allowance, and Moz's own definition is that one row typically corresponds to one object in the response for methods that return multiple objects, or one field in the response for methods that return a single flat object or report. That gives four cost shapes across this connector, and every node states which one applies to it. One row per call covers the single-answer nodes such as Moz: Fetch Site Metrics and Moz: Get Link Status. One row per returned result covers every list node, which means a list node's page size is its price. Moz: Fetch Keyword Metrics bills one row per returned value up to four, with nulls not charged. And Moz: Fetch Link Intersects bills rows equal to the number of targets for every result, which makes it the most expensive node here. Four nodes are free: Moz: Fetch Quota, Moz: Fetch Current Index ID, Moz: Get Top Ranking Domains and Moz: Get Top Ranking Pages. Moz publishes a tiered pricing model with pay-as-you-go overages on some tiers, and the current tier table is on Moz's own pricing page rather than here, because it is the kind of number that should be read from the source.

Does this work with Claude, ChatGPT, or other AI assistants?

Yes. Mattering exposes these nodes over MCP, an open protocol for connecting AI assistants to tools, so any MCP-capable assistant, including Claude and ChatGPT, can connect to your workspace and use them. Matter, the assistant built into the platform, uses the same nodes natively. The point worth making on a metered API is that the same nodes also run unattended on a schedule or a trigger, which is where the free quota check and the free index check earn their place: a guardrail only helps if it runs when nobody is watching.

Is there a free Moz API tier, and can I test with it?

Yes, there is a free tier, and no, it will not carry a working automation. Moz's authentication guide documents a free tier for users who want to test the API before purchasing, offering 50 rows of data per month, with a valid non-expired credit card required to sign up and no charge unless you choose to upgrade. Moz's help documentation adds that free access allows one request every ten seconds. Fifty rows is enough to prove a credential and run a handful of single-row calls. It is not enough to run a report: one call to Moz: List Links at the default page size of 25 spends half of the month. The Moz API is also a separate subscription from Moz Pro and Moz Local, so a Moz Pro seat on its own does not give you API access.

Does testing the connection spend quota?

No, and that was the deciding factor in how the test was built. Mattering probes the credential with Moz's quota lookup method, which Moz's own documentation states does not require any quota and presents as the call you can run at no cost. On a metered API, a connection test that charges you every time somebody clicks it would be a defect rather than a detail. A passing test proves the URL, the header name, that the token is valid, and that the account holds a Data API subscription, and it reports the remaining row allowance as a bonus. What it does not prove is that a paid method will succeed: a token whose monthly rows are exhausted still passes this probe and then fails on the first real call.

Moz publishes its own MCP server. Why use this connector?

Moz does, and for chatting with your Moz data it is a reasonable answer. Moz documents two endpoints, one for Moz Data and one for Moz Local, each requiring a valid subscription to the corresponding product, both published in Claude's built-in connector directory, and both authenticated with either OAuth in the client or the same API token from the dashboard. Moz also states that the server is in beta and that access is limited to account owners today, with expansion to seated users in progress. This connector answers a different question. It is 31 typed nodes that run on a schedule or a trigger with nobody present, chain with every other connector in one workflow, hold the token once in a shared workspace store rather than in an individual's client config, and carry their row cost in the node description so a workflow can be reviewed for spend before it runs. Neither replaces the other, and the honest summary is that theirs is a chat surface and this is an automation surface.

What Moz operations are supported?

31 nodes: 30 read nodes and one action. Site metrics cover Moz: Fetch Site Metrics, Moz: Fetch Multiple Site Metrics, Moz: Fetch Brand Authority, Moz: Fetch Site Metrics Distributions and Moz: Fetch Site Metric Histories. Links cover Moz: List Links, Moz: Filter Links by Anchor Text, Moz: Filter Links by Domain, Moz: Fetch Link Intersects, Moz: Get Link Status, Moz: Get Multiple Link Status and Moz: List Anchor Texts. Linking domains cover Moz: List Linking Domains, Moz: List Recently Gained Linking Domains and Moz: List Recently Lost Linking Domains. Pages and redirects cover Moz: List Top Pages and Moz: Get Final Redirect Target. Rankings cover Moz: List Ranking Keywords and Moz: Count Ranking Keywords. Keyword research covers Moz: Fetch Keyword Metrics, Moz: Fetch Keyword Difficulty, Moz: Fetch Keyword Volume, Moz: Fetch Keyword Organic CTR, Moz: Fetch Keyword Priority, Moz: Fetch Keyword Search Intent and Moz: List Related Keywords. Reference and account cover Moz: Get Top Ranking Domains, Moz: Get Top Ranking Pages, Moz: Fetch Quota and Moz: Fetch Current Index ID. The single action is Moz: Custom API Method.

Does it cover Moz Local as well as the Data API?

Not as typed nodes, on purpose. Moz Local is a different product with its own subscription, and Moz splits it the same way in its own MCP guide, one endpoint per product. It is also the only write surface in the Moz API, including methods that cancel or close a listing, and shipping destructive nodes for a product this connector cannot test would be the wrong first move. Moz Local methods remain reachable through Moz: Custom API Method, which calls any Moz method by name, and that node's description says plainly that some Local methods write and that its cost is whatever the named method costs. Using it against Moz Local needs a Moz Local subscription in addition to the Data API one.

What are Moz's rate limits?

Moz documents an error-driven throttle rather than a flat ceiling. Its error handling guide states that making more than 50 requests in a five minute period that return a 4xx caps the account at 50 requests per five minutes, until 30 seconds pass without a 4xx. In other words, a workflow that is failing fast gets slowed down, and a workflow that is succeeding is not the target of that rule. On the free tier Moz's help documentation also documents one request every ten seconds. Errors come back as a JSON-RPC error object carrying an HTTP status and a data key that names the offending part of the request, which is specific enough to branch on rather than merely log.

31 nodes waiting on the canvas

Put Moz to work inside your workspace

Connect once, then use every node in visual automations and AI agent workflows across the Mattering platform.

All product names, logos, and brands are property of their respective owners and are used for identification purposes only. Use of these names, logos, and brands does not imply endorsement or affiliation. Moz is a product of its respective owner; Mattering is an independent platform.

Start Free Explore our Products