Jube
jube.io
- Free plan
- Linux
- Web
- API
- Self-hosted
What Jube is
Jube is an open-source platform for real-time suspicious-transaction detection, AML case management, and fraud prevention. Its monitoring combines thresholds, velocity and aggregation checks, sanctions screening, and supervised and unsupervised machine learning. Rules and model outputs can work together, and the platform supports real-time scoring, custom rules, TTL counters, and suppression.
Interfaces can operate synchronously or asynchronously over HTTP and AMQP, including hybrid modes. Redis holds real-time state, PostgreSQL provides durable storage and audit logs, and RabbitMQ handles inbound and outbound events over AMQP. Case management covers AML, fraud, and compliance, with investigator dashboards, document versioning, automatic escalation, and audit trails. Jube runs in Docker and Kubernetes, supports multiple tenants, and can be deployed on customer infrastructure or cloud tenancy. It is free under AGPLv3 with all features available; modified core software run as a network-accessible service must make its modified source available to users of that service. The Docker Compose setup is for evaluation and development; production deployments are directed to clustered deployment documentation.
Who it suits
Jube is aimed at compliance teams, financial institutions, fintechs, and organizations adopting AML or fraud detection. It may suit teams able to deploy and operate containerized software and account for the AGPLv3 terms.
Good to know, and what to watch
- Free AGPLv3 software with all features available
- Combines rules, sanctions screening, and machine learning
- AML, fraud, and compliance case management
- Supports Docker, Kubernetes, and multiple tenants
- Can run on customer infrastructure or cloud tenancy
- AGPLv3 requires sharing modified core source in specified network-service use
- Docker Compose setup is for evaluation and development only
- Growth support covers one pipeline under roughly 100,000 monthly transactions
Plans
Every plan and priceJube
Jube software
Sandbox support
Growth support
Free plan
Getting started
- Open the Jube website and review the self-hosted AGPLv3 software
- Deploy the containerized platform with Docker or Kubernetes
- Configure PostgreSQL, Redis, and RabbitMQ integrations
- Connect transaction flows through HTTP, AMQP, or hybrid interfaces
- Choose optional sandbox, growth, scale, enterprise support, or training services
Limits to check
The free Jube software includes all features, but it is self-hosted under AGPLv3. The $350-per-month Sandbox support plan is for non-production use only and provides async responses within 48 hours. Growth support covers one live production pipeline under roughly 100,000 transactions per month.
The Club's review
Jube is an open-source platform for real-time suspicious-transaction detection, AML case management, and fraud prevention. It suits financial institutions, fintechs, and compliance teams able to operate self-hosted infrastructure. The main appeal is a broad monitoring and investigation stack; the main commitment is running and governing it as production software under AGPLv3.
Overview
Jube brings transaction scoring, rules, machine learning, sanctions screening, and investigation workflows into one system. It can run on an organisation’s infrastructure or cloud tenancy, including in the organisation’s jurisdiction, so data can remain under its control. That is a strong fit for regulated operators with deployment expertise and data-residency requirements, but less suitable for teams seeking a fully managed service.
Key features
Detection and explainability
The engine combines threshold, velocity, aggregation, TTL-counter, suppression, and sanctions checks with supervised risk models and unsupervised anomaly detection. Its exhaustive adaptation approach tries neural-network structures and variables, allowing explicit controls and model-led signals to work together. This breadth can cover varied AML and fraud programmes, although designing, tuning, and validating those controls remains a specialist responsibility.
Jube says Bayesian Network and decision-tree outputs can be explained to regulators and describes calibration and validation against BASEL practice. That emphasis is valuable where investigators and auditors must understand why a transaction was flagged, rather than receiving an opaque score.
Case management
AML, fraud, and compliance streams share investigator dashboards, document versioning, automatic escalation, and audit trails. Keeping cases beside monitoring can shorten the path from an alert to documented review and preserve an evidential record. The coverage is broader than a card-fraud-only workflow and therefore better suited to organisations with several compliance operations.
Integration and deployment
Jube supports synchronous and asynchronous processing through HTTP, AMQP, or hybrid interfaces. PostgreSQL provides durable storage and audit logs, Redis holds real-time state, and RabbitMQ handles inbound and outbound events. Container images support Docker and Kubernetes, while multi-tenancy helps organisations separate customers or business units.
The software is self-hosted and fully open source under AGPLv3, with all features included. If modified core software is offered as a network-accessible service, the maker says those modifications must be made available to service users. The Docker Compose setup is intended for evaluation and development; production deployments should use the clustered deployment documentation.
Pricing
The Jube software plans cost nothing. “Jube” is self-hosted AGPLv3 software, and “Jube software” also includes all features under AGPLv3. There is no free trial because the software itself is free.
| Plan | Price | What it includes |
|---|---|---|
| Sandbox support | $350 a month | Non-production sandbox; async response within 48 hours |
| Growth support | $950 a month | One live production pipeline, under roughly 100,000 transactions a month; next-business-day response |
| Scale support | $1,590 a month | Multi-region pipelines; same-day, high-priority routing |
| Enterprise support | $2,400 a month ($28,800 a year) | Mission-critical infrastructure, 24/7 priority emergency SLA, advanced analytics consulting; 12-month initial term |
Sandbox is the sensible choice for non-production evaluation. Growth fits a single smaller production pipeline, while Scale is aimed at multi-region operation. Enterprise is for mission-critical deployments that need emergency coverage, security patching, and consulting; its annual commitment matters. The maker gives no transaction-volume cap for the other support tiers.
Public training in Kuala Lumpur costs $1,825, paid once per participant for three days, up to six participants, with materials, workshops, lunch, refreshments, and up to six weeks of remote support. Private training starts at $6,000, paid once, with on-site pricing based on location and scope. Support is async-first over WhatsApp, with Loom and documentation responses tracked as tickets; enterprise callout response is guaranteed 06:00–18:00 GMT Monday to Friday, with best-efforts coverage outside those hours.
Platforms
Jube is offered through an API, on Linux, on the web, and as self-hosted software. Docker and Kubernetes support make container orchestration the natural production route.
Who it's for
Choose Jube if your financial institution, fintech, or compliance operation needs real-time monitoring and case management under one deployment you control. It is particularly relevant where jurisdictional data control, auditable decisions, and multiple AML or fraud streams outweigh the operational work of running clustered infrastructure. Small teams wanting a hosted, low-maintenance detector should look elsewhere.
Pros and cons
- Pro: Rules, sanctions checks, supervised models, and unsupervised detection provide several complementary ways to identify risk.
- Pro: Investigator dashboards, escalation, versioning, and audit trails connect alert handling with compliance evidence.
- Pro: Self-hosting and multi-tenancy support control over data location and organisational separation.
- Con: Production requires clustered deployment guidance rather than the simpler Docker Compose evaluation setup.
- Con: AGPLv3 obligations matter if you modify the core and expose it as a network service.
- Con: Growth support covers only one pipeline under roughly 100,000 monthly transactions, and support is asynchronous outside enterprise emergency arrangements.
Alternatives
Payment Fraud Detection Software is a broader category for comparing specialist options. Pick FraudLabs Pro instead when a freemium, hosted service with a 500-query monthly Micro plan and a free trial is preferable. IPQualityScore Fraud Prevention suits teams wanting a freemium API with Android, iOS, web, or self-hosted access and a 1,000-lookups-per-month free tier.
Stripe Radar is the alternative for businesses already using Stripe standard payments that want included AI card-fraud and card-testing prevention. Choose Ojuri for MIT-licensed self-hosted software with no per-call fees or rate limits. MaxMind minFraud is better when pay-as-you-go scoring, custom inputs, and custom rules are the priority.
SAS Visual Statistics is an enterprise-scale statistical modelling option on SAS Viya. Tazama is another free, open-source choice, with hosting, implementation, and support requiring resources or a contract. Ravelin is a paid, tailored platform for organisations seeking a customised setup.
Verdict
Jube is a strong choice for regulated financial organisations that want open-source, self-hosted monitoring and investigation with explainable model outputs and flexible event interfaces. Choose it for control, breadth, and integrated case work; look elsewhere if you need a managed service, a lighter deployment, or support without AGPLv3 and infrastructure responsibilities.
The verdict
Jube combines transaction monitoring with case management and supports deployment on infrastructure controlled by the customer. The software is free, while support and training are separately priced; production deployment requires the clustered deployment guidance.
Questions people ask
Is Jube free?
Yes. The Jube and Jube software plans are listed as free, with AGPLv3 licensing and all features available.
Where can Jube run?
Jube supports API, Linux, self-hosted, and web use. It is containerized for Docker and Kubernetes and can run on your infrastructure or cloud tenancy.
Which systems does Jube integrate with?
PostgreSQL provides durable storage and audit logs, Redis stores real-time state, and RabbitMQ supports inbound and outbound AMQP events.
What does the AGPLv3 licence require?
The maker says modified core software operated as a network-accessible service must make its modified source available to service users.
What paid support plans are available?
Sandbox support costs $350 a month, Growth $950 a month, Scale $1,590 a month, and Enterprise $2,400 a month.
What the maker publishes
- Case management
- Case management includes AML, fraud, and compliance streams, investigator dashboards, document versioning, automatic escalation, and audit trails. source
- Data control
- The maker says deployment on the customer’s infrastructure and in its jurisdiction means data does not have to leave the customer’s control. source
- Deployment
- Jube is containerized for Docker and Kubernetes, supports multi-tenancy, and can run on an organisation’s own infrastructure or cloud tenancy. source
- Detection
- Its detection combines thresholds, velocity and aggregation checks, sanctions screening, and supervised and unsupervised machine learning. source
- Enterprise coverage
- Enterprise support includes security patching and a callout response guaranteed 06:00–18:00 GMT Monday to Friday, with best-endeavours coverage outside those hours. source
- Explainability
- The site says Bayesian Networks and decision-tree rule output can be explained to regulators and describes calibration and validation against BASEL practice. source
- Headquarters
- Nottingham, England, UK source
- Integrations
- The product page lists PostgreSQL for durable storage and audit logs, Redis for real-time state, and RabbitMQ via AMQP for inbound and outbound events. source
- Intended users
- The maker describes Jube as intended for compliance teams, financial institutions, fintechs, and organisations adopting it for AML or fraud detection. source
- Interfaces
- Jube supports synchronous and asynchronous interfaces via HTTP, AMQP, and hybrid modes. source
- License
- The site says Jube is fully open source under AGPLv3 and that all features are available. source
- License limit
- Jube is offered under AGPLv3; the maker says modified core software run as a network-accessible service must make its modified source available to users of that service. source
- Machine learning
- It combines unsupervised anomaly detection, supervised risk models, and exhaustive adaptation that tries neural-network structures and variables. source
- Production setup
- The GitHub README says its Docker Compose setup is for evaluation and development only and directs production deployments to clustered deployment documentation. source
- Purpose
- Jube is an open-source platform for real-time suspicious-transaction detection, AML case management, and fraud prevention. source
- Rules and sanctions
- The rule engine supports thresholds, velocity and aggregation checks, sanctions screening, TTL counters, and suppression, integrated with ML outputs. source
- Security support
- Enterprise support includes patching credible, documented vulnerabilities, prioritized by severity and exploitability. source
- Support
- Support is async-first over WhatsApp, with Loom video and documentation responses tracked against a permanent record. source
- Support limitation
- The Growth support tier covers one production pipeline under roughly 100,000 transactions a month; the page gives no transaction-volume cap for the other tiers. source
- Support model
- Support is async-first over WhatsApp, with requests tracked as tickets and answered with Loom videos and documentation. source
- Training
- The maker offers hands-on training and implementation, with up to six weeks of remote support in the listed training offers. source
- Transaction monitoring
- The monitoring engine supports synchronous and asynchronous interfaces via HTTP, AMQP, and hybrid modes, with Redis state and PostgreSQL storage and audit logs. source
Checked against the maker's pages on 8 Oct 2026.