Claude Sumsub connectorWhat Claude can do in your Sumsub account.
The Claude Sumsub connector exposes 18 tools: 6 in read, 3 in write and 9 that no source describes. Claude reads applicants, their timeline, transactions and verification levels, and can create an applicant and a verification link, within your Sumsub role.
Verified Trustpilot reviews · AI, automation & growth agency
What changes when Claude can work in Sumsub
Sumsub's directory description promises compliance work in plain language without clicking through the Dashboard. With the connector, the question goes to Claude in plain words. It pulls the applicant, the events behind the last review or the flagged transactions, and can start an onboarding, always within what your Sumsub role already allows.
Answering a reviewer's question fast. Asked why an applicant was last reviewed, Claude fetches the record with applicant_get and reads the event history through applicant_timeline_get.
Triage on transaction monitoring. transaction_list filters an applicant's transactions, for example those of the last 7 days that hit a rule, and transaction_get opens the details of one.
Onboarding from the chat. Claude creates an individual applicant at a chosen level with applicant_create_individual, then produces the link to send them through verification_link_create.
What it will not do, or not visibly: half of the 18 tools on the directory sheet are described by no source, so this page makes no claim about them. Nothing runs by itself either: no alert reaches Claude when a check fails, and it acts only while it answers you. Sumsub's directory description also mentions dossiers and regulatory reports, but no tool is documented for those. The Claude connectors hub lists other finance connectors, and the Integrations page covers automation platforms.
Five words before you connect
The Claude vocabulary behind this connector, in one minute.
- Connector
- The link you set up once between Claude and an account you already hold, so Claude can work in it while it answers you.
- Tool
- One named thing the connector lets Claude do on that account. Claude chooses the ones it needs, and the directory sheet lists them by name.
- Authorization
- Sumsub's sign-in screen, where you hand Claude the access it will use afterwards. It is granted once per person and can be taken back.
- Approval
- The confirmation Claude waits for before it goes ahead with something that changes an account, shown in the chat at the right moment.
- MCP
- The shared standard connectors are built on: it is what lets an assistant such as Claude talk to an outside service like Sumsub.
Connecting Sumsub to Claude in three steps
- 01
Find Sumsub in Claude
Open Claude's settings, go to Customize then Connectors, and search for Sumsub. On a Team or Enterprise workspace, an Owner or Primary Owner enables the connector first so that each member can then authenticate with their own account.
- 02
Start the connection
Click Connect on its row and sign in to Sumsub in the window the service opens. If the connection breaks later, use Disconnect and connect again: that usually restores it.
- 03
Read the authorization screen
Look at Sumsub's authorization screen before you accept. It belongs to Sumsub, not to Claude, and it decides what the access covers. You can also remove that access later from the Sumsub side.
The 18 Sumsub tools, sorted by what they do
Sumsub gives Claude 18 tools: 6 that read your account, 3 that change something in it, and 9 no official source describes.
Sumsub's MCP page documents nine of them through worked examples; the other nine appear on the directory sheet with nothing more than a name. Names stay as Claude shows them.
- 6 read
- 3 write
- 9 not documented
What Claude reads (6)
6 toolsSix tools that look up applicants, their events, transactions and verification levels.
verification_level_list
Retrieves the verification levels configured on your Sumsub account, so Claude can tell you which ones exist before looking closer at any of them.
verification_level_get
Inspects one or more selected verification levels in detail, which lets Claude lay two configurations side by side and explain where they differ.
applicant_get
Retrieves the data Sumsub holds on one applicant, identified by its ID, so Claude can answer questions about that person or company from the record itself.
applicant_timeline_get
Reads the sequence of events recorded for an applicant, which is how Claude can explain what happened before and after a given review.
transaction_get
Opens one transaction and returns its details, so Claude can discuss a specific payment flagged during monitoring instead of a whole list.
transaction_list
Lists transactions with filters, for instance those of one applicant over a recent period that hit a monitoring rule, giving Claude a working set to triage.
What Claude can create or send (3)
3 toolsThree tools that create an applicant, a verification link or a message to Sumsub. No source describes a confirmation of their own: the general rule below applies.
applicant_create_individual
Approval: see the ruleCreates an individual applicant in Sumsub at a verification level you name, the first step Sumsub shows for onboarding someone from a conversation.
verification_link_create
Approval: see the ruleGenerates a verification link for an applicant, the link that person then follows to complete their checks, as Sumsub's onboarding example describes.
feedback
Approval: see the rulePasses your suggestion to the Sumsub team through the built-in feedback tool, when a tool, parameter or workflow you need is missing from the connector.
Not documented (9)
9 toolsListed by name on the directory sheet, described by no official source. This page does not guess what they do.
how_to_sumsub
Sumsub's MCP page walks through several request examples, and this tool appears in none of them; the directory sheet only prints its name. With no official description, this page does not say what it returns. Like any call, it stays within your Sumsub role permissions and falls under the general approval rule.
questionnaire_list
No sentence from Sumsub or Anthropic describes this one: it shows up on the directory sheet as a bare name and nowhere in the publisher's connector documentation. Its effect is therefore not stated here. The default approval rule in the block further down is what applies when Claude calls it.
questionnaire_get
Absent from the examples Sumsub publishes for its server, this tool has no documented behaviour. Guessing from the name is exactly what this page refuses to do. What can be said: Sumsub ties every agent action to the permissions of your role, and Claude's general approval rule covers the call.
vasp_list
Listed on the directory sheet, undocumented everywhere else we could read, including Sumsub's MCP server page. Without a source, its output stays unknown and nothing is assumed about whether it reads or alters data. If Claude reaches for it, the general rule on approvals set out below applies.
applicant_status_get
Sumsub's own status example for an applicant relies on two other tools, not this one, and no official text explains it. Its behaviour is therefore left undocumented on this page. Your role permissions in Sumsub still bound any call, and the approval rule described further down governs it.
applicant_list
The publisher documents nothing about this tool beyond the fact that it sits on the directory sheet. No description, no example and no parameter have been published. This page stays silent on its effect and points to the general approval rule, which covers it like every other call.
applicant_aml_case_get
Although the directory description mentions screening cases in general, no official source ties any sentence to this specific tool. It is treated as undocumented, with no claim here about what it opens or returns. The default approval rule and your Sumsub role permissions are the only documented frame around it.
applicant_fraud_network_list
The directory overview speaks of fraud networks in broad terms, yet no source explains what this particular tool does. This page keeps it in the undocumented group rather than infer a capability. Any call goes through Claude's general approval rule and stays limited to what your Sumsub role allows.
applicant_create_company
Unlike the individual onboarding tool, which Sumsub's example names, this one is never described by any source, publisher or Anthropic. Its name is not evidence, so no effect is stated here. Should Claude call it, the general approval rule of the block below and your Sumsub role permissions apply.
What Claude asks before acting
By default, Claude stops and asks for your go-ahead before any action it takes on an account in your place. With three tools that create records or send a message, that default matters.
On Team and Enterprise workspaces, owners decide whether a member may let some actions through without being asked every time, and they can bound what a connector may do for the whole organisation, keeping reads and closing writes for instance. Nobody bypasses that from their own account. Sumsub adds its own layer: the agent can only perform what your Sumsub role permissions allow, and the role itself must carry the Use MCP server permission.
Which plans include it
None of the 819 sheets in the official directory displays availability by plan. Connector by connector, the answer is published nowhere: it is a real gap in the catalogue.
What is published is the general rule: remote connectors are open to all users on Claude, Cowork, Claude Desktop and mobile. On Team and Enterprise, an Owner or Primary Owner enables the connector for the organisation before each member signs in. For the current state of this one, check its sheet in the official directory. Note that only Sumsub documents it; no Claude help article covers it.
Where the Claude Sumsub connector stops
A connector is not an automation. Claude calls these tools while it answers you: nothing fires when an applicant fails a check or a transaction trips a rule.
The tool list is an observed floor, not a promise: an administrator can open actions that no public sheet shows. The partner badge in the directory is not a security audit, and Anthropic says on every sheet that it does not choose the tools a publisher exposes and cannot guarantee they behave as announced. Connect only what comes from a publisher you trust, here Sumsub, whose MCP page is the only documentation of this connector.
Need help connecting Sumsub to Claude?
A person reads every message.