Mentingo is an open-source, self-hosted learning management system (LMS) with a built-in AI mentor, built for corporate L&D, employee onboarding and compliance training.
MIT-licensed, white-label and multi-tenant - run it on your own infrastructure, brand it as your own, and keep every learner record inside your organisation.

Live demo · Quick start · Deploy on AWS · Features · Security & compliance · FAQ · Docs · Commercial support

Course catalog - continue learning, top courses, one click in	Lesson view - distraction-free content with video
AI Mentor - real-time role-play, scored automatically	Assessment engine - scenario-based questions, not just recall
Admin dashboard - deadlines, to-do list, AI mentor practice	Recertification - expiring certificates, signature and validity period

Enterprise L&D and HR teams running onboarding, product and mandatory training for hundreds or thousands of employees, who need one platform instead of ten tools - and who need to prove, on demand, who was trained and when.

Compliance, security and risk owners who have to deliver recurring, evidenced training (cyber security, safety, regulatory) and need learner data to stay in a jurisdiction and an infrastructure they control.

Product and engineering teams building an EdTech or HRTech product, or embedding training into an existing platform. MIT licence, typed OpenAPI client, multi-tenancy and white-labelling are in the box - no per-seat licence to resell around.

Runs locally on macOS, Linux or Windows. Full guide: docs/development-setup.md. Environment variable reference: docs/configuration.md.

Prerequisites: Node.js >= 22.15 (see .tool-versions), pnpm 10, Docker + Docker Compose, Caddy 2.8.4.

git clone https://github.com/Selleo/mentingo.git
cd mentingo
pnpm setup:unix    # Windows: pnpm setup:win
pnpm dev

setup verifies tooling, configures HTTPS via Caddy, installs dependencies, writes .env files, starts the Docker services (PostgreSQL + pgvector, Redis, MinIO, Mailhog), runs migrations and seeds a three-tenant demo environment.

Then open https://tenant1.lms.localhost and sign in:

API and Swagger docs: https://tenant1.lms.localhost/api · Mail catcher: https://mailbox.lms.localhost

Deploying to a server: docs/deployment.md (worked example on Hetzner Cloud + AWS Route 53/ECR).

Do not want to build anything? Mentingo is listed on the AWS Marketplace as a container image with a CloudFormation template that provisions ECS, RDS and ElastiCache for you.

• No software charge. The listing is free with no end date - you pay only for the AWS resources the stack creates.
• Runs in your own AWS account, in the region you choose. Learner data never leaves it.
• Subscribe, launch the stack, point a domain at it.

• Practise conversations instead of watching slides. A voice and chat AI mentor runs real-time role-play for sales, compliance and customer-support scenarios, and scores the attempt automatically.
• Turn the documents you already have into a course. The AI course architect converts knowledge bases, policies and internal documentation into structured, interactive courses in minutes.
• Grade open-ended answers without an L&D queue. Behavioural and problem-solving tasks are analysed automatically and returned with actionable feedback.
• Keep the AI accountable. Every model call is traced with Langfuse, so you can inspect cost, latency and what the mentor actually said.

• Ship it as your own product. Custom domains, logos, colours and styles - no Mentingo branding required anywhere in the learner experience.
• One deployment, many organisations. Multi-tenancy is enforced in PostgreSQL with Row-Level Security, not only in application code - fit for holding structures, group companies and B2B training vendors.
• Turn infrastructure on without a developer. A central admin hub enables corporate SSO (Microsoft 365, Google Workspace, Slack) and your AI provider key from the UI.

• Mirror the org chart you actually have. Departments, branches and teams as groups, with course access inherited automatically.
• Mandatory training that renews itself. Set expiry intervals for compliance, safety or cyber-security courses; re-enrolment and reminder schedules run on their own.
• An interface people open voluntarily. A clean, consumer-grade UI built for completion rates, not for admin convenience.
• Assess more than recall. Behavioural scenarios, short and long-form text analysis, dynamic quizzes and daily streaks.
• See what training is doing. Engagement, consistency and outcome analytics per learner, group and course.

• One place for the material that is not a course. Central repository for documentation, playbooks and assets.
• Keep hybrid teams aligned. Built-in company blog, announcements and an internal FAQ.

Licences verified from each project's LICENSE file in September 2026. Commercial pricing and feature tiers change - check the vendors' current pricing pages before making a decision.

Choose Mentingo if you want a modern, AI-first LMS you can host yourself, brand as your own and extend in TypeScript, without a per-seat licence and without a copyleft obligation.

Choose something else if you need a decades-old plugin ecosystem for academic teaching (Moodle), MOOC-scale public course delivery (Open edX), or a fully managed enterprise suite and you are comfortable with per-user pricing (Docebo, TalentLMS).

Component diagram, multi-tenancy model and how the AI mentor pipeline fits together: docs/architecture.md. Environment variable reference: docs/configuration.md.

Apps

Packages

What it costs to run

The repository ships a k6 performance suite (pnpm perf:load, perf:stress, perf:spike, perf:soak) so you can size your own deployment before committing to it.

Mentingo is built for training that has to be evidenced to an auditor, not only delivered.

Tenant and data isolation

• Multi-tenancy enforced at the database level with PostgreSQL Row-Level Security.
• The application connects with a least-privilege runtime role; schema migrations use a separate administrative role.
• Each tenant is served on its own origin, with CORS restricted per tenant.

Authentication and access

• Corporate SSO over OAuth 2.0: Microsoft 365 (incl. Microsoft Graph), Google Workspace, Slack.
• Email and password with bcrypt hashing, JWT access and refresh tokens.
• Time-based one-time password (TOTP) two-factor authentication.
• Role-based access: admin, content creator, learner; group-based course access inheritance.
• API rate limiting and request throttling.

Data handling

• Self-hosted: learner data stays in your database, your object storage and your region. You choose the cloud and the jurisdiction - an EU-only deployment is a configuration choice, not a paid tier.
• Files are served through pre-signed, expiring URLs; video through Bunny Stream signed tokens.
• An application master key is required to run the platform and is used to protect sensitive stored values.
• AI: bring your own key. You configure the AI provider and its API key. No prompt or learner content is routed through Selleo infrastructure in a self-hosted deployment.

Evidence and observability

• Recurring recertification with expiry dates, automated re-enrolment and notification schedules.
• Per-learner completion and engagement records, exportable to CSV/XLSX.
• Application monitoring with Sentry and OpenTelemetry; every AI call traced in Langfuse.

Interoperability

• SCORM 1.2 export, so content authored in Mentingo can be handed to another LMS - and so your organisation is not locked in.

Found a vulnerability? Please read SECURITY.md - do not open a public issue.

• Typed API client. The API is documented with OpenAPI/Swagger at /api; pnpm generate:client regenerates the typed client used by the web app - and by anything else you build.
• Bulk operations. Import and export users and results as CSV or XLSX.
• Integration origins. A dedicated CORS allow-list (INTEGRATION_CORS_ORIGINS) for embedding Mentingo endpoints in your own products.
• Your own AI provider. Model and key are configuration, not a fork.
• White-label. Domains, logo, colours and styles per tenant.
• Localisation. Six UI languages shipped; translations live in apps/web/app/locales/<lang>/translation.json - adding one is a pull request, not a licence upgrade.

Contributions welcome - see CONTRIBUTING.md and the Discussions board.

Mentingo is an open-source, self-hosted learning management system (LMS) with a built-in AI mentor. It is built for corporate learning and development - employee onboarding, product and sales enablement, and mandatory compliance training - rather than for academic course delivery.

The software is free under the MIT licence. You can run it, modify it, rebrand it and even resell it without paying a licence fee. What you pay for is the infrastructure you run it on, and optionally a managed hosting or implementation contract with Selleo.

Yes - self-hosting is the default. You need Node.js 22, PostgreSQL 16 with the pgvector extension, Redis and any S3-compatible object storage. docs/deployment.md walks through a production deployment on Hetzner Cloud with AWS for DNS and container registry, but nothing ties the platform to those providers.

Yes, in two ways. Mentingo is listed on the AWS Marketplace as a container image with a CloudFormation template that provisions ECS, RDS and ElastiCache in your own AWS account - the listing itself is free of charge, you pay only for the AWS resources, and it lands on your existing AWS bill instead of going through a new vendor procurement cycle. Alternatively, Selleo will host and operate Mentingo for you on a flat tiered subscription.

Moodle is a mature, plugin-driven platform designed for education, licensed under GPL v3 and written in PHP. Mentingo is a TypeScript platform designed for corporate L&D, licensed under MIT, with an AI mentor, multi-tenancy and white-labelling in the core rather than in plugins. If you need a large academic plugin ecosystem, choose Moodle. If you need a modern, brandable corporate training platform you can extend yourself, choose Mentingo.

It covers the same job - corporate training, compliance and analytics with AI support - without a per-user subscription and without giving up control of your data. The trade-off is that you operate it yourself, or pay someone to operate it for you. Docebo is a managed suite with a longer feature list and a matching bill.

Mentingo can export courses as SCORM 1.2 packages, so content you build here can be handed to another LMS.

You configure the provider and supply your own API key - OpenAI is supported out of the box through the Vercel AI SDK. In a self-hosted deployment, prompts and learner content go from your server to the provider you chose and nowhere else; nothing is routed through Selleo. Every model call is traced locally with Langfuse, so you can audit what the mentor said, what it cost and how long it took.

Yes. Courses can be given expiry intervals with automated re-enrolment and reminder schedules, completion records are kept per learner and exportable, and each tenant's data is isolated at the database level with PostgreSQL Row-Level Security. Whether that satisfies a specific regulation is a question for your compliance team and your deployment, not for the software alone.

Yes - OAuth 2.0 with Microsoft 365, Google Workspace and Slack, plus email and password with TOTP two-factor authentication.

Yes. Multi-tenancy is built in and enforced in PostgreSQL with Row-Level Security, with per-tenant domains and branding. This is what makes Mentingo usable by holding structures, group companies and B2B training vendors.

That depends on the number of active learners, how much video you serve and how heavily you use AI. The repository ships a k6 performance suite (pnpm perf:load, perf:stress, perf:spike, perf:soak) so you can size a deployment against your own numbers before committing to it.

Yes. The MIT licence permits commercial use, modification and redistribution, and does not require you to publish your changes. Domains, logo, colours and styles are configurable per tenant.

English, Polish, German, Spanish, French and Czech. Translations live in apps/web/app/locales/<lang>/translation.json - adding a language is a pull request.

Selleo, a product engineering company. Mentingo is a real product with a managed offering, not an abandoned demo repository - see CHANGELOG.md for the release history.

• Shipped changes: CHANGELOG.md · GitHub Releases
• Planned work: Issues

Mentingo is free to self-host under the MIT licence, and it is maintained by Selleo - a product engineering team that has been building learning platforms since 2005.

Talk to us →

• Discussions - questions, ideas, show and tell
• Issues - bugs and feature requests
• CONTRIBUTING.md - branch, commit and pull request conventions
• CODE_OF_CONDUCT.md

If Mentingo is useful to you, a ⭐ helps other L&D and engineering teams find it.

MIT - see LICENSE. You may run, modify, white-label and resell Mentingo, including commercially, without a copyleft obligation to publish your changes.

Built and maintained by Selleo.