The Whole Platform, Open Source

Yes. The whole platform is open source at github.com/Aelira-AI/aelira-core — the engine under AGPL-3.0, the command-line client under MIT. You can download, install, modify, and run Aelira for free forever. No usage limits, no time restrictions, no feature gates, no hidden fees. We offer paid support from $799/month if you want installation help, supported upgrades, model guidance, and ongoing maintenance. The software itself remains free.

The engine is AGPL-3.0; the command-line client in cli/ is MIT so you can embed and script against it freely. Both halves are open source and both are in the same public repository — the split is about how you can reuse each part, not about holding anything back. AGPL on the engine keeps the platform open: if you modify it and offer it to others over a network, those changes go back to the community. Universities running it internally are unaffected — you can modify it freely for your own institution without publishing anything. MIT on the CLI removes that obligation entirely, so it can go into your own build pipelines, internal tools, or commercial products with no copyleft reach.

Accessibility is a human right, not a profit center. Universities shouldn't have to commit to expensive, quote-based licenses for closed-source software just to remediate their documents. Our business model: we give away the software for free and sell support, hosting, and professional services. This aligns our incentives - we make money by helping you succeed, not by locking you into expensive licenses.

Self-hosted (free): • You clone the repo and run it yourself • You provide infrastructure (servers, storage, backups) • Community support via GitHub Issues • No support SLA Paid support (from $799/month): • Guided installation and upgrade planning • Documented response targets • Supported release and model guidance • Quarterly reviews on institution plans The software is identical. You are paying for support or managed infrastructure, not a separate feature-locked edition.

v0.9.5 is the current stable release, published August 23, 2026. The functionality is complete and the engine is what runs our own cloud. Operational hardening — upgrade paths, long-run performance under heavy batch load, and deployment edge cases — continues transparently. We track that work in public issues at github.com/Aelira-AI/aelira-core, so you can see exactly what's known before you commit to a deployment. If you hit something we haven't, open an issue.

Yes, under both licences. You can: • Add custom features for your university • Integrate with your existing systems • Rebrand the UI • Fix bugs yourself • Contribute improvements back (appreciated but not required) The one obligation AGPL-3.0 adds on the engine: if you offer your modified version to others over a network, publish your changes. Internal institutional use carries no such requirement, and the MIT-licensed CLI carries no such requirement at all. If you modify the platform and want us to maintain your fork, that falls under our professional services ($200-$350/hr).

Minimum requirements: • 4-8 CPU cores • 8-16GB RAM • 50-100GB storage (SSD recommended) • Docker and Docker Compose Recommended for production: • 8-16 cores (better AI performance) • 16-32GB RAM (parallel processing) • 200-500GB storage (scan history) • PostgreSQL database • Redis cache • A reverse proxy for TLS You can host on: department servers, AWS, Azure, GCP, DigitalOcean, Hetzner, OVH, or any VPS provider.

No. The docker compose quickstart brings up everything: the API, the dashboard, PostgreSQL, and Redis. AI is configurable. You can point Aelira at a hosted model, or run it fully locally with Ollama so nothing leaves your network - the repo documents both paths in docs/deployment/local-ai-models.md. No machine learning background required, just basic Docker and Linux knowledge.

Yes. We welcome contributions: • Bug fixes • New features (especially higher ed-specific) • Documentation improvements • Translations • Test coverage • Performance optimizations See CONTRIBUTING.md in the repository for guidelines. All contributors are credited in CHANGELOG.md.

We offer professional services for custom development: • API integration with existing systems • Custom workflow development • Migration from legacy tools (YuJa, Ally, ReadSpeaker) • Training for IT staff (1-5 days) • On-site implementation Rates: $200-$350/hr depending on complexity. Or hire your own developers - both licences allow it.