canonical: https://jentic.com/apis/fraudlabspro.com/fraudlabs-pro

# FraudLabs Pro API

FraudLabs Pro screens online transactions for chargeback risk before merchants commit to fulfilment. The version 2 API exposes five endpoints covering real-time order screening, asynchronous order-result retrieval, post-transaction feedback (approved, rejected, chargeback) for model improvement, plus an SMS verification flow for step-up authentication. It returns a numeric fraud score, distance and proxy signals, and a recommended action so merchants can auto-approve, review, or reject.

## For AI agents

Score an order for fraud risk in real time, send post-transaction feedback to improve the model, and run SMS step-up verification when needed. Authentication is an API key in the 'key' query or body parameter.

## Scope

Does not handle payment processing, refunds, or chargeback dispute filings - use for fraud scoring, post-transaction feedback, and SMS verification only.

## Capabilities

- Score an order for fraud risk via `/order/screen` with billing, shipping, IP, and payment fields
- Retrieve a previously screened order's verdict and signals via `/order/result`
- Send post-transaction feedback (approved, rejected, chargeback) via `/order/feedback` to improve the model
- Trigger an SMS verification code to a customer phone via `/verification/send`
- Confirm an SMS verification code via `/verification/result` for step-up authentication

## Use cases

### E-Commerce Checkout Risk Scoring

An online store calls `/order/screen` at checkout with the cart, billing address, shipping address, IP, and payment method. FraudLabs Pro returns a numeric score plus an action recommendation (approve, review, reject) within a few hundred milliseconds, letting the store auto-fulfil low-risk orders and route high-risk ones to manual review.

Example prompt: POST to `/order/screen` with the order payload and the API key in the 'key' parameter, then branch on the recommended action returned by the API

### Closed-Loop Feedback for Model Tuning

After the merchant settles or charges back an order, the integration calls `/order/feedback` so FraudLabs Pro learns whether its earlier verdict was correct. Sending feedback is what keeps the per-merchant fraud model accurate over time and is the single highest-leverage thing an integrator can do.

Example prompt: POST `/order/feedback` with the order id and status (APPROVED, REJECTED, or CHARGEBACK) once the bank settles or disputes the transaction

### SMS Step-Up Verification

When the screen verdict is REVIEW, the merchant triggers `/verification/send` to push an OTP to the customer's phone, then confirms the code through `/verification/result.` A successful OTP softens the risk signal and lets the order proceed without manual review.

Example prompt: POST `/verification/send` with the customer phone, then on user input POST `/verification/result` with the otp_code and check the response state

### AI Agent Risk Triage

An agent monitoring a checkout queue uses Jentic to score orders, send feedback, and request SMS verification when the score is borderline. The agent never holds the FraudLabs Pro key directly and surfaces a plain-language summary of why each order was approved or flagged.

Example prompt: Search Jentic for 'score an order for fraud risk', load the `/order/screen` schema, execute against each pending order, and post a triage summary back to operations

## Key endpoints

| Method | Path | Description |
| --- | --- | --- |
| POST | `/order/screen` | Score an order for fraud risk |
| GET | `/order/result` | Retrieve a previously screened order's verdict |
| POST | `/order/feedback` | Send post-transaction feedback (APPROVED, REJECTED, CHARGEBACK) |
| POST | `/verification/send` | Trigger an SMS verification code |
| POST | `/verification/result` | Confirm an SMS verification code |

## Key resources

- **Order** — Screen orders, retrieve previous results, and submit post-transaction feedback
- **SMS Verification** — Send OTPs and verify them for step-up authentication

## Why Jentic

- **Setup:** Wiring the FraudLabs Pro API by hand means appending your API key to each request and building your own retry and error handling across its fraud-scoring, feedback, and SMS verification endpoints. Through Jentic you install once, import FraudLabs Pro from the API Directory, store the key once, and your agent calls it.
- **Permission scoping:** FraudLabs Pro carries the order and verification details in the request body rather than the URL path, so scoping is operations-only: you limit the agent to the operations it needs, such as screening an order or reading a result, and leave out sending SMS codes or posting feedback unless you add them. Every operation the agent can run is one you chose.
- **Credential handling:** Your FraudLabs Pro API key is stored once, encrypted, by your own Jentic One instance and injected at execution time. It never enters the agent's prompt, logs, or context.
- **Discovery method:** Agents search Jentic by intent such as 'score an order for fraud risk' or 'send chargeback feedback', and Jentic returns the matching FraudLabs Pro operation with its input schema so the agent does not need to read the developer docs.

## Related APIs

- **FraudLabs Pro Fraud Detection (v1)** — Earlier v1 fraud detection API with the same screening concept under `/v1/order` paths
- **Sift** — Sift is a higher-end fraud and trust platform with broader event types and ML model controls
- **Fingerprint** — Fingerprint provides a stable visitor identifier that strengthens FraudLabs Pro's IP and device signals

## FAQ

### What authentication does the FraudLabs Pro API use?

FraudLabs Pro uses an API key passed as the 'key' query or body parameter on every call. Through Jentic, the key is stored encrypted in the vault and substituted at execution time, so the raw key never enters the agent's context.

### Can I score an order for fraud risk with the FraudLabs Pro API?

Yes. POST to `/order/screen` with the order payload - billing address, shipping address, customer IP, payment method, item list, and amounts. The response includes a numeric fraud score, a recommended action, and risk signals such as IP geolocation and proxy detection.

### What are the rate limits for the FraudLabs Pro API?

FraudLabs Pro applies per-plan monthly transaction quotas rather than per-second rate limits, with the free Micro plan allowing a few hundred screens per month and paid plans scaling into the millions. Quota is enforced on `/order/screen` calls; feedback and result retrieval do not count.

### How do I send chargeback feedback through Jentic?

Search Jentic for 'send fraud feedback', load the `/order/feedback` schema, and execute it with the order_id from the original screen call plus the status set to CHARGEBACK. This retrains the per-merchant fraud model so future screens are more accurate.

### Is the FraudLabs Pro API free?

FraudLabs Pro offers a free Micro plan with a small monthly screen allowance, plus paid Mini, Small, Medium, and Large plans that scale by transaction volume and include faster support. SMS verification is billed per message in addition to the screen quota.

### How do I run SMS step-up verification on a borderline order?

When `/order/screen` returns a REVIEW recommendation, POST `/verification/send` with the customer's phone number to deliver an OTP, then POST `/verification/result` with the user-entered code. A successful verification softens the risk score and lets the order proceed without human review.

### Can I limit what my agent is allowed to do with the FraudLabs Pro API?

Yes. Because you self-host Jentic One, your own rules decide which FraudLabs Pro operations the agent may call, and it can only run the ones you grant. FraudLabs Pro carries order and verification details in the request body rather than the URL, so scoping is operations-only: you can allow the agent to screen an order via `/order/screen` and read a verdict via `/order/result` while leaving out sending SMS codes through `/verification/send` or posting feedback via `/order/feedback.` Your API key is injected only for the operations you approve, so every call the agent can make is one you chose.