Ojuri

Ojuri is a self-hosted fraud detection system for payment operators that need to make accept, review, or decline decisions on transactions. It is best suited to fintechs, payment gateways, and mobile-money platforms with the capacity to run their own software. Its no-per-call-fee, MIT-licensed plan is attractive, but deployment and ongoing operation remain the operator’s responsibility.

Overview

Ojuri brings transaction scoring, fraud-pattern analysis, investigations, and analyst tools into one system. Its omnichannel support and API make it relevant to operators working across payment channels, while self-hosting keeps transaction data within the operator’s infrastructure for scoring, training, and investigation.

That control has a practical cost: the quickstart requires Docker and Node 20. Investigation reports can use a self-hosted language model, but enabling it requires downloading about 7.6 GB of weights. Self-hosting can help with data-residency requirements, including GDPR, NDPR, and CBN; it is not, by itself, a compliance certification.

Key features

Transaction scoring and relationship analysis

The Real-Time Detection Agent scores payments on the authorization path using an XGBoost model served through ONNX Runtime. That design is aimed at teams that need fraud decisions within payment authorization. The Pattern Analysis Agent tracks relationships among accounts, devices, and payees over time to identify possible fraud rings and mule networks, adding a view of connected activity alongside individual transaction scores.

Auditable decisions and analyst workflow

Ojuri records decision reason codes and the data the model saw in an audit log, giving teams a way to trace why a transaction received its outcome. Sentinel adds a review queue, decision stream, model registry, and analyst overrides that can become training data. Together, those tools suit teams that want analysts to feed operational decisions back into the fraud workflow.

Model learning and investigations

The Model Learning Agent monitors drift, retrains on verified labels, and gates model promotion on evaluation checks. This offers a structured path for updating models, though it still depends on verified labels and a team able to operate the system. The Fraud Investigation Agent creates reports for declined transactions using a self-hosted language model and runs separately from live scoring, so investigation work is distinct from the authorization path.

Integration and data controls

Applications can send requests to the POST /v1/predict API; the Node SDK supports typed requests, retries, and webhook signature verification. The documented controls also include API keys, dashboard accounts and roles, signed webhooks, and idempotency keys for safe retries. Ojuri says transaction data stays inside the operator’s infrastructure for scoring, training, and investigation, which makes the deployment model relevant to teams with data-boundary requirements.

Pricing

Ojuri: free. The plan is MIT licensed, with no per-call fees and no rate limits. It includes custom rules, machine-learning models, and device fingerprinting, with omnichannel support. That makes the software cost predictable for operators with variable transaction volumes, but “free” does not remove the work or infrastructure involved in self-hosting.

Platforms

Ojuri supports API and web access, self-hosted deployment, and Windows. The quickstart uses Docker Compose and requires Docker and Node 20; separate Windows setup instructions are provided. Teams should account for those runtime requirements, especially if they enable the optional investigation model and its approximately 7.6 GB download.

Who it's for

Ojuri is a strong fit for fintechs, payment gateways, and mobile-money platforms that want control over transaction data and a configurable fraud workflow spanning live scoring, analyst review, and model learning. It is less suitable for teams seeking a hosted service that avoids deployment and maintenance work, or for organizations treating self-hosting as a substitute for compliance certification.

Pros and cons

• Pros: No per-call fees or rate limits make the MIT-licensed plan appealing for operators whose transaction volume may grow or fluctuate.
• Pros: Scoring, relationship analysis, analyst review, and model learning sit within one self-hosted workflow, supporting teams that want to manage these stages together.
• Pros: Decision reason codes and recorded model inputs help analysts trace outcomes in the audit log.
• Cons: Operators must run the system themselves; the quickstart requires Docker and Node 20.
• Cons: The optional investigation model adds a substantial download of about 7.6 GB, which teams must plan for if they want that capability.
• Cons: Self-hosting can support data residency but does not certify compliance with GDPR, NDPR, CBN, or other requirements.

Alternatives

Stripe Radar is worth considering for businesses already on standard payments pricing that want AI fraud prevention for card payments and card testing included at no additional charge through Radar Lite. Its free plan is tied to that standard payments pricing, unlike Ojuri’s self-hosted MIT-licensed model.

FraudLabs Pro may suit teams that prefer a query-based free tier: its Micro plan allows 500 queries per month, sales under 25K USD per month, five validation rules, no grouping rules, no sub-accounts, and 50 custom blacklists. That cap makes it a different fit from Ojuri’s no-rate-limit plan.

IPQualityScore Fraud Prevention is an option for teams that want a defined lookup allowance and custom-rule quota: its Free plan includes 10 custom rules, 1,000 lookups per month, 35 lookups per day, and three CSV uploads per day. Jube is another free, self-hosted option, with software under AGPLv3 and all features available.

Tazama offers free open-source software with no software or license charge, though hosting, implementation, and support require resources or a contract. Ravelin has a customized platform with tailored setup and contact-based pricing. Vesta is a paid alternative. Cybersource is a paid global payment gateway with pricing available by contacting the provider.

For a broader comparison, see Payment Fraud Detection Software.

Verdict

Choose Ojuri if you are a payment operator that wants to keep transaction data within your own infrastructure and combine scoring, investigation, and analyst workflows without per-call fees or rate limits. Its main advantage is that breadth and control in a free, MIT-licensed package. Look elsewhere if your team cannot take on self-hosted deployment and upkeep, or needs compliance certification rather than infrastructure-level data control.