PlugWiki
NoResponse logo

NoResponse MCP Server: Setup for Claude, Cursor and Other Apps

OfficialRemoteLocal
Getnpm@noresponse/mcpnoresponsehq/mcp

Lets Claude or Cursor check signups, logins and purchases for fraud, and manage accounts and block lists in NoResponse.

This is the NoResponse company's own server for its risk API. The assistant can assess an email, phone, username or IP address, read an account's risk decision, review linked accounts and record what you did about them.

It runs on your computer with an API key, or you can point an app at the hosted address and send the key in an Authorization header. It is for teams that already use NoResponse. The README lists 30 tools, but the code registers 96, so many more reach beyond fraud checks to reviews, exports and team management.

Good for

  • Fraud and trust teams who want an assistant to look up account risk
  • Developers wiring NoResponse into a signup or login flow
  • Support staff handling account erasure requests

Works in

14 should work

Tap an app to see how to set it up and any known problems.

Install NoResponse MCP

Claude Code

  1. 1

    Copy this prompt.

    Add the NoResponse MCP server to Claude Code. Run this command:
    
    claude mcp add --transport http noresponse https://mcp.noresponse.org/mcp --header "Authorization: Bearer <NORESPONSE_API_KEY>"
    
    Ask me for NORESPONSE_API_KEY and only put the values in that config, nowhere else.
    
    Then check that it loaded and tell me if I need to sign in or restart.
  2. 2

    Paste it into Claude Code. It shows the change and asks before running it.

Requires NoResponse API key.

Read-only mode: Add --tool-annotations readOnly after start to mount only the tools marked read-only. A key without the delete, management and webhooks scopes also keeps those tools unusable.

Adapted from the README: the README shows the local command with start and the hosted address with a Bearer header; this config puts the same values in this app's format. Hosted: put your key in an Authorization header (Bearer). The README does not name this app.

Built from the package data we hold. Not yet tested by us.

Tools

  • Checks a signup, login or purchase for fraud risk

    Sends the email, phone, username, name or IP address to NoResponse and asks for a risk assessment of an account creation, login or purchase. A new request ID is added if you give none.

    Try: “Assess a signup from jane@example.com at IP 203.0.113.7.”

    Write · create_assessment

  • Reads the latest result of one assessment

    Fetches the newest version of a single assessment by its ID, including its risk and recommendation.

    Try: “Show me the latest result for assessment as_123.”

    Read · get_assessment

  • Waits for a pending assessment to finish

    Checks a pending assessment again and again until it settles or a timeout passes (25 seconds by default, 55 at most).

    Try: “Wait for assessment as_123 to finish and tell me the decision.”

    Read · wait_for_assessment

  • Lists an account's recent assessments

    Pages through the assessments of one account that have not expired, newest first.

    Try: “List the recent assessments for account acct_42.”

    Read · query_assessments

  • Shows an account's current risk and decision

    Returns the account's risk score and its decision for each flow (allow, step_up, review, block or held), plus the assessment behind it. Needs a key with the history scope.

    Try: “What is the current risk decision for account acct_42?”

    Read · get_account

  • Lists an account's assessments by account ID

    Same listing as the query tool, with the account ID in the request path. Needs the history scope.

    Try: “Page through the assessments for acct_42.”

    Read · list_account_assessments

  • Finds accounts that share an email, phone or username

    Lists other accounts in the same environment that submitted an email, phone or username this account also submitted, with the shared values. Needs the history scope, or investigate with evidence.

    Try: “Which other accounts share details with acct_42?”

    Read · list_linked_accounts

  • Lists accounts whose risk has changed

    Returns accounts whose risk changed, oldest change first. Pass the cursor back to poll for more, or ask for only accounts that still need action. Needs the history scope.

    Try: “Which accounts changed risk since my last check and still need action?”

    Read · list_changed_accounts

  • Records what you did to an account

    Stores the action you took in your identity provider for one flow at one risk version, as applied or failed. Repeating the call returns the stored record. Needs the history scope.

    Try: “Record that we blocked acct_42 at login for risk version 3.”

    Write · record_enforcement

  • Reports what you later learned about an account

    Tells NoResponse the real result for an assessed account, such as confirmed abuse or legitimate.

    Try: “Mark assessment as_123 as confirmed abuse.”

    Write · record_outcome

  • Deletes one assessment and its evidence

    Removes a single assessment with its evidence and outcomes, then scores the account again. Needs the delete scope and is marked destructive.

    Try: “Delete assessment as_123.”

    Delete · delete_assessment

  • Erases an account and its whole history

    Erases every assessment of the account with its evidence and outcomes, for an erasure request, and sends an account.deleted webhook if anything was removed. Needs the history scope and is marked destructive.

    Try: “Erase acct_42 and all its history.”

    Delete · delete_account

Read from the source code

Permissions

high risk · 5 things to watch

High risk

The assistant can erase accounts and assessments, revoke API keys and change block lists in your NoResponse workspace. Use a key without the delete and management scopes, or start the server with the read-only filter.

What it can do

Watch out for

  • It can erase accounts and delete assessments

    delete_account erases every assessment of an account and delete_assessment removes one, both marked destructive. Give the server a key without the delete scope, or mount read-only tools. github.com

  • It can revoke keys and remove team members

    revoke_key and the member and invite tools are marked destructive and change who can use your NoResponse workspace. Use a key without the management scope. github.com

  • It can change who gets blocked

    Settings and list tools change thresholds and the allow and block lists, and the matching accounts are decided again. Check the assistant's plan before you approve a change. github.com

  • Network mode listens on every interface

    Starting it with the SSE transport or the serve command binds 0.0.0.0, not just localhost. Keep the default stdio mode, or put a firewall in front. github.com

  • Identifiers go to NoResponse

    Assessing a signup sends the email, phone, username, name or IP to the NoResponse API. Tell your users in your privacy notice. github.com

Security advisories

No published security advisories as of Nov 1, 2026.

Technical details
Turning on read-onlystart --tool-annotations readOnly mounts only tools marked read-only (src/mcp-server/cli/start/command.ts). A key without the delete, management and webhooks scopes also keeps those tools unusable.
Talks toNoResponse API (identifiers you assess are sent there). mcp.noresponse.org when you use the hosted address
Sign-in and keysAPI key in the NORESPONSE_API_KEY environment variable, or an Authorization header for the hosted address. The README says the server stores nothing.
Scopes
  • history

  • delete

  • management

  • webhooks

  • integrations

Questions

Is the NoResponse MCP server official?

Yes. It comes from the NoResponse organisation on GitHub and is built from the same API description as its SDKs.

Do I need an API key?

Yes. You need a NoResponse API key. Some tools also need a key with extra scopes, such as history, delete, management or webhooks.

Can I run it read-only?

Yes. Start the local server with the tool-annotations option set to readOnly and only tools marked read-only are mounted. A key without the delete scope also blocks deletes.

Is it local or hosted?

Both. You can run it on your computer over stdio with npx, or use the hosted address with your key in an Authorization header. The hosted address is https://mcp.noresponse.org/mcp.

Can it delete data?

Yes. It can erase accounts, delete assessments, revoke API keys, remove members and delete webhooks, integrations and list entries. Remove those scopes from the key you give it.

Why does the tool list show only some tools?

The README lists 30 tools, but the code registers 96, including reviews, exports, policy evaluations and team tools. The list on this page covers the 30 the README documents.

Does it work in Claude Desktop?

Yes. The project ships a desktop extension file that Claude Desktop opens and asks for your API key. You can also use the npx config in the install box.

Information

updated Nov 2, 2026

Publisher
Official
Package
@noresponse/mcp
Registry
npm
Transport
stdio, Streamable HTTP
Auth
API key
Version
0.11.7
License
Apache-2.0
Last commit
Oct 31, 2026
Open issues
0

WebsiteGitHubnpmChangelogIssues

Sources

Docs, issue trackers and discussions this page is built from. Issues, quotes and findings above link to their own source.

Show all 5
  1. github.com · github.com
  2. github.com · github.com
  3. github.com · github.com
  4. github.com · github.com
  5. github.com · github.com

Facts from the official MCP registry, GitHub, npm, PyPI and the Docker catalog. Updated Nov 2, 2026.