| Website | https://www.mozilla.ai/ |
| Agent Platform | https://www.mozilla.ai/product/agent-platform |
| Choice-First Stack | https://www.mozilla.ai/open-tools/choice-first-stack |
We had a great couple of days at Data Makers Fest 2026 in Porto from May 5–6.
Toto (@totostache) presented a live demo of https://otari.ai, while Davide Eynard (@mala) and Kevin Gallacio joined him throughout the event speaking with data engineers, analysts, and AI teams about on-prem deployments, LLM gateways, and open-source AI infrastructure.
Thanks to everyone who stopped by the booth!
Benchmarks matter, but they’re not the whole story.
Mozilla.ai CEO John Dickerson will discuss why ease of use may be the key to broader open source AI adoption.
Most people don’t need the latest frontier model. They need tools that are simple to install, maintain, and use in daily workflows.
He’ll also cover Mozilla.ai’s “choice-first stack” and building open source AI tools that feel practical and approachable.
See you there May 12–14: https://link.mozilla.ai/ai-council-2026
Octonous is now in open beta.
One of the biggest things we learned during the closed beta:
people wanted more visibility and control before trusting automation.
That feedback led to memory, approval flows, clearer logs, expanded integrations, and model choice.
Start automating your tasks: https://octonous.com
On Your Terms is a https://Mozilla.ai series exploring how AI should be built and used.
The first piece is based on a conversation with John Dickerson, CEO of https://Mozilla.ai, on sovereign AI.
It spans:
• Nation-state
• Company
• Community
• Individual
For companies, it means control over their AI stack and avoiding reliance on a single provider.
Full post: https://link.mozilla.ai/sovereign-ai
Own your AI agent: running open-source agents on your terms.
David de la Iglesia Castro and Davide Eynard (@mala) will walk through how to build and run open-source AI agents in practice.
🗓️ May 7–9 & 15–16 (online)
Join here: https://link.mozilla.ai/oxml
We’re heading to Data Makers Fest next week!
We’ll be demoing https://otari.ai and will have a https://Mozilla.ai booth.
One gateway across LLM providers with key management, budget controls, and usage tracking built in.
Your prompts stay private.
Come by and see what we're building.
Guardrails in multiple languages require more than translation.
They must detect threats across context and culture.
Alinia brings that into any-guardrail.
Learn more: https://link.mozilla.ai/alinia-any-guardrail
The newest integration with any-guardrail: Alinia AI, whose security models are specifically built to detect threats like prompt injection, data exfiltration, and policy violations by understanding the cultural and linguistic nuances of multilingual AI interactions.
Running AI locally still takes more setup than people expect.
llamafile keeps it simple.
One file.
Run it.
You’re using the model.
No Python. No Docker.
Try it here: https://link.mozilla.ai/llamafile-repo
Last week we hosted an Octonous workshop in Berlin 🇩🇪
People came in with real work and spent time building and testing ideas.
Thanks to everyone who joined and shared their workflows.
And thank you to Taisia Bekbulatova for capturing the event!
We’ll be running more of these soon.
The more “clever” the packaging, the harder it is to understand what’s inside.
Encoderfile leans into a more “dull” format you can inspect, validate, and reason about.
That matters in production.
Encoder models power most NLP in production, but deploying them still means dragging along Python runtimes and dependencies. Encoderfile introduces a single executable with an appended payload and a format that can be inspected and understood.