Sift (Chargeback) MCP. Manage risk scores and apply decisions in chat.
Gemini Works with every AI agent you already use
…and any MCP-compatible client
Just plug in your AI agents and start using Vinkius.
Sift (Chargeback) is an MCP Server for managing fraud prevention and chargebacks. It connects your AI agent to Sift's risk engine, allowing you to retrieve real-time user fraud scores, track custom events, report disputes, and apply manual decisions—all without leaving your chat interface.
What your AI agents can do
Apply user decision
Applies a manual decision (e.g., block_user) to an account.
Get user fraud labels
Retrieves labels (like $bad or $good) assigned to a specific user.
Get user fraud score
Returns the current fraud risk score for a given user ID.
Retrieves a numerical fraud score for a specified user ID.
Gets specific labels (e.g., $bad, $good) Sift has assigned to a user based on their activity.
Applies manual actions or automated decisions (like blocking or approving) against an account.
Submits a formal chargeback report to Sift, updating the user's risk profile.
Lists all past decisions and labels applied to a single user ID for review.
Records general behavioral events (like logins or transactions) into Sift's system for ML training.
Ask AI about this MCP
Supported MCP Clients
Gemini Waiting for input…
Sift (Chargeback): 8 Tools for Fraud Monitoring
These tools let your agent interact directly with Sift's API. You can check scores, log events, apply decisions, and audit history all through natural conversation.
019d756capply user decision
Applies a manual decision (e.g., block_user) to an account.
019d756cget user fraud labels
Retrieves labels (like $bad or $good) assigned to a specific user.
019d756cget user fraud score
Returns the current fraud risk score for a given user ID.
019d756clist sift decisions
Shows all predefined actions or decisions available in Sift.
019d756clist sift workflows
Retrieves a list of configured fraud prevention workflows within your account.
019d756clist user decision history
Lists every decision and label applied to an account over time.
019d756creport sift chargeback
Reports a specific chargeback event, updating the user's risk profile immediately.
019d756ctrack sift event
Logs general behavioral events (like login or transaction) to refine Sift’s machine learning model.
Choose How to Get Started
Build a custom MCP for your own tools, or connect a ready-made integration from our catalog.
Build Your Own
Turn any API into an MCP. Import a spec, define Agent Skills, or deploy with MCPFusion.
- Import from OpenAPI, Swagger, or YAML specs
- Create Agent Skills with progressive disclosure
- Deploy to edge with MCPFusion framework
- Built in DLP, auth, and compliance on every call
- Real time usage dashboard and cost metering
- Publish to catalog or keep private
Make Your AI Do More
Start with Sift (Chargeback), then connect any of our 4,700+ other servers whenever your AI needs more. One click, no limits.
- Use this MCP plus 4,700+ others, all in one place
- Add new capabilities to your AI anytime you want
- Every connection is secured and compliant automatically
- Track usage and costs across all your servers
- Works with Claude, ChatGPT, Cursor, and more
- New servers added to the catalog every week
What you can do with this MCP connector
Listen up. This Sift (Chargeback) MCP Server gives your AI client direct control over fraud prevention and chargebacks. You don't gotta stare at a dashboard to manage risk anymore; you just talk to the server, and it handles the heavy lifting.
This isn't some wrapper that just sends data off. It exposes specific tools that let your agent speak straight to Sift’s core API. Here’s what you can do with it:
Checking a User's Risk Status
To see if a user is risky, the get_user_fraud_score tool pulls their current numerical fraud score. That gives you an immediate number to work with. You can also run get_user_fraud_labels to check what specific labels—like $bad or $good—Sift has already slapped on that user based on past activity. If you wanna see all the actions available, run list_sift_decisions; it lists every predefined decision point Sift uses.
Logging and Tracking Behavior
Every action a user takes matters for fraud modeling. Use track_sift_event to log general behavioral data—think logins, transactions, or page views—which feeds back into Sift's machine learning model. If a chargeback happens, you don’t wait for manual reporting; the report_sift_chargeback tool submits that formal report right away, updating the user's risk profile instantly.
Making Decisions and Adjusting Accounts
When you know the risk score, you gotta act. The apply_user_decision tool lets you manually apply an action to an account—you can block a user or greenlight an order—based on your judgment. You can also check what kinds of fraud prevention workflows are active in your setup by calling list_sift_workflows, which shows all the configured rules.
Auditing Everything That Happened
You gotta keep records, right? The list_user_decision_history tool pulls a complete timeline, showing every single decision and label that’s ever been applied to an account. This is your audit trail. You can also run list_sift_decisions again just to see the full menu of predefined actions available in Sift.
Here's how it works with your agent:
- You hook up this server and feed it your Sift REST API Key and Account ID.
- When you ask your AI client, 'What’s the risk for user X?' your agent recognizes that it needs to run a tool. It calls
get_user_fraud_scorewith User X's ID. - The MCP Server hits the Sift API, grabs the raw score data, and feeds it straight back to your agent. Your agent then takes that cold data and spits out an easy answer for you.
It’s simple: talk naturally, and your agent does all the heavy lifting across fraud risk management without you ever leaving your chat window.
How Sift (Chargeback) MCP Works
- 1 Subscribe to the server and input your Sift REST API Key and Account ID.
- 2 Your AI agent calls a tool function (e.g.,
get_user_fraud_score) with necessary parameters like a User ID. - 3 The MCP Server executes the call against the live Sift API and passes the resulting data payload back to your agent.
The bottom line is, it lets you use established fraud APIs as if they were simple chat commands.
Who Is Sift (Chargeback) MCP For?
Risk Analysts, Trust & Safety Managers, and E-commerce Operations teams need this. If your job involves manually checking dashboards or clicking through logs to determine why an order was rejected—you're here. This tool lets you automate the investigative part of fraud review.
Runs queries on user history (list_user_decision_history) and checks scores (get_user_fraud_score) to identify patterns in chargeback activity.
Applies immediate, manual decisions (e.g., blocking a user) directly from the chat interface using apply_user_decision without opening the Sift dashboard.
Verifies transaction risk and reports chargebacks (report_sift_chargeback) immediately when suspicious activity is flagged, minimizing financial loss.
What Changes When You Connect
- Real-time risk assessment: Need to know if a user is safe? Use
get_user_fraud_scoreto pull the current score instantly, letting you decide whether to approve or deny an order before it happens. - Instant dispute action: Don't wait for tickets. With
apply_user_decision, your agent can block a fraudulent account or manually accept a legitimate claim right from your conversation window. - Complete audit trail: Tracking fraud decisions is critical. Use
list_user_decision_historyto pull a full, auditable timeline of every action taken against an account ID. - Efficient incident handling: When you spot suspicious activity, call
report_sift_chargeback. This immediately logs the event and updates the user's risk profile in Sift. - Data enrichment for ML: You can refine your fraud model without writing code. Just use
track_sift_eventto log custom activities like 'login' or 'purchase_attempt'. - Workflow transparency: Use
list_sift_workflowsto see exactly what automated rules are running in Sift, helping you understand why a user might be flagged.
Real-World Use Cases
Investigating Account Takeover (ATO)
Problem: A customer reports suspicious activity. Manually checking the fraud dashboard takes time. Action: Your agent first calls get_user_fraud_score to get a baseline risk number, then runs list_user_decision_history to see if past decisions flagged anything. Result: The agent tells you they are high-risk because of repeated label assignments and immediate actions.
Processing a Chargeback Dispute
Problem: A large chargeback hits the system, requiring an immediate report. Action: You tell your agent to 'Report this chargeback for order #123.' The agent uses report_sift_chargeback, attaching the reason and details. Result: Sift instantly updates the user's risk score, informing you if follow-up action is needed.
Onboarding a New Client
Problem: You need to ensure all new users are tracked correctly from day one. Action: Your agent runs track_sift_event every time a user successfully logs in or completes a signup form. Result: Sift's ML model gets continuous, clean data on the user's initial behavior, improving future risk predictions.
Reviewing Compliance Logs
Problem: A compliance officer needs to prove that specific users were manually reviewed and approved. Action: The agent uses list_user_decision_history for the target user ID, filtering by date range. Result: You get a clean list of all manual decisions (e.g., 'manual_review', 'approved') with timestamps ready for audit.
The Tradeoffs
Assuming risk is static
A developer checks the fraud score once and assumes it's good, then approves a large transaction without logging anything. This leaves an audit gap.
→
Always use get_user_fraud_score for the check, but also call track_sift_event immediately after approval to log the transaction context. Then, if needed, follow up with list_user_decision_history.
Forgetting event logging
A user signs up and logs in via a webhook, but nobody calls an API tool to record this critical behavior.
→
Every key behavioral change must be logged. Use track_sift_event with parameters like 'user_login' or 'signup_complete'. This feeds the ML model.
Treating decisions as magic
A manager says, 'Just block this user.' without knowing what specific decision code Sift requires.
→
First, run list_sift_decisions to see the exact available action names (e.g., 'block_user'). Then, use apply_user_decision with that precise name.
When It Fits, When It Doesn't
Use this server if your primary bottleneck is transforming raw Sift API data into conversational decisions. You need to check scores (get_user_fraud_score), log history (list_user_decision_history), or execute immediate actions (apply_user_decision) without leaving the chat client.
Don't use this if you are building a massive, multi-service event pipeline that requires guaranteeing transactional consistency across multiple external data sources. For that, an asynchronous message queue (like Kafka) is better. If your need is simple lookups and immediate actions based on Sift’s existing logic, however, this MCP server is the right fit.
Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by Sift. All third-party trademarks, logos, and brand names are the property of their respective owners. Their use on this website is strictly for informational purposes to identify service compatibility and interoperability.
VINKIUS INFRASTRUCTURE
Cloud Hosted
Managed infra
V8 Isolated
Sandboxed per request
Zero-Trust Proxy
No stored credentials
DLP Enforced
Policy on every call
GDPR Compliant
EU data residency
Token Compression
~60% cost reduction
Works with Claude, ChatGPT, Cursor, and more
The Model Context Protocol standardizes how applications expose capabilities to LLMs. Instead of operating in isolation, your AI gains direct access to external platforms, live data, and real-world actions through secure, standardized connections.
This server provides 8 capabilities that interface natively with Claude, ChatGPT, Cursor, and any MCP client. No middleware. No custom integration required.
Available Capabilities
Reviewing a user's risk profile used to be a multi-tab dashboard nightmare.
Today, checking a single customer's history means opening the main dashboard. You click the 'Scores' tab for the score, then navigate to the 'Disputes' section to see if they reported a chargeback, and finally find the 'History' log to confirm who approved them. It’s three separate pages just to get one picture.
With this MCP server, you ask your agent: 'What's the full risk profile for user X?' The agent runs `get_user_fraud_score`, pulls labels via `get_user_fraud_labels`, and checks `list_user_decision_history`—all in one response. You get the answer without clicking anything.
Sift (Chargeback) MCP Server: Decision & Reporting
Before, applying a decision meant logging into the platform and manually selecting an option from a dropdown menu. If you needed to block them, that was one button; if you needed to review history, it was another screen entirely.
Now, your agent handles it all. You simply tell the client: 'Block this user because they charged back.' The server runs `apply_user_decision` and logs the action automatically. It's direct, immediate control.
Common Questions About Sift (Chargeback) MCP
How do I check the fraud score using get_user_fraud_score? +
Just ask your agent for the 'fraud score for user ABC'. The tool calls Sift and returns a real-time percentage risk score, telling you if they are high or low risk.
What is report_sift_chargeback used for? +
Use report_sift_chargeback when you need to officially log a chargeback event. It submits the dispute details to Sift, which immediately updates the user's risk profile.
Can I track custom events with track_sift_event? +
Yes. You can tell your agent to 'Log a successful login for this user.' The tool calls track_sift_event, logging that specific behavior into Sift's system.
How do I see past fraud decisions with list_user_decision_history? +
Ask your agent to 'Show me the decision history for user XYZ.' The tool runs list_user_decision_history and gives you a clean timeline of every label or action ever applied.
How does using `apply_user_decision` actually affect a user's account? +
It immediately updates the user's status within Sift. For example, calling it with 'block_user' prevents them from transacting until an administrator manually lifts the block. This is how you enforce policy decisions directly via your AI client.
How do I check my configured fraud prevention workflows using `list_sift_workflows`? +
This tool retrieves a list of all active fraud rules and workflow pipelines set up in Sift. You can review these to understand the criteria that automatically score or label users, helping you debug unexpected risk changes.
What's the difference between getting a score and using `get_user_fraud_labels`? +
The fraud score gives a continuous percentage of calculated risk. In contrast, get_user_fraud_labels returns discrete, categorical labels (like '$bad' or '$good'). Labels represent Sift's final classification based on the raw score and rules.
Before I apply an action, how do I find all possible actions using `list_sift_decisions`? +
list_sift_decisions pulls a reference list of every decision Sift supports. This is crucial for ensuring you use the correct string or command name when building automated workflows with your agent.
Can I block a fraudulent user through the agent? +
Yes! Use the apply_user_decision tool with the user's ID and the appropriate decision ID (like block_user). The action will be applied in Sift immediately.
How do I check the risk score for a specific customer? +
Use the get_user_fraud_score tool with the customer's unique ID. Your agent will fetch the latest risk score generated by Sift's machine learning models.
Where do I find my Sift API Key and Account ID? +
Log in to your Sift Dashboard and navigate to Settings -> API Keys. You will find your REST API Key and Account ID there.
Use it with your favorite AI tools
Connect this server to Cursor, Claude, VS Code, and more.
More in this category
AdGuard Home
Manage your AdGuard Home instance — control protection, DNS settings, and filtering rules directly from any AI agent.
Green Web Check
Universal sustainability intelligence — check if websites are hosted on green energy via AI.
Azure Functions Invoke
This MCP does exactly one thing: it invokes a single Azure Function. That's its only function, and nothing else. Incredible for letting your AI execute secure serverless compute.
You might also like
YouTube
Search for videos, retrieve detailed statistics, and audit channel performance on YouTube — the world's leading video sharing platform.
Indiegogo Crowdfunding API
Access crowdfunding data — audit campaigns, funding, and categories via AI.
Zillow Alternative
Access real estate data, Zestimates, and property details directly from Zillow — search addresses, get valuations, and analyze market trends.