← Feed

Posts

• View post

  Between job sites, client calls, and paperwork, tradespeople are juggling a lot every day. Clawbolt helps keep work moving by turning your messaging app into an AI assistant for the business side of the job. Draft invoices. Organize job photos. Manage schedules. Follow up with customers. Works with tools like QuickBooks, CompanyCam, Google Calendar, and more. Sign up for early access: https://link.mozilla.ai/clawbolt-home

• View post

  Ever wondered how any-llm knows which model provider you’re using? Nathan Brake explains how any-llm parses the provider and model ID, then loads the correct SDK behind the scenes at runtime. It’s one of the things that makes switching between providers feel seamless. Rewatch the Open Source Friday recording: https://link.mozilla.ai/osf-any-llm

• View post

  If engineers can’t fully reason about every line of code, reliability has to live somewhere else. Observability, rollbacks, isolation, and fast recovery become essential. Less about perfect code. More about systems that tolerate failure. If you missed our piece on this shift: https://link.mozilla.ai/owning-code-ai

• View post

  The interesting shift in AI software may not be agents clicking buttons inside existing apps. It may be the work itself moving into structured systems agents and humans can both operate on directly. This piece explores what that could mean for docs, dashboards, workflows, CRMs, and presentation software. Read the full piece: https://link.mozilla.ai/the-interface-is-no-longer-the-product

• View post

  Less than one week until SAO. https://mozilla.ai is helping co-organize the Supporting Our AI Overlords workshop together with Bauplan, MongoDB, Datadog, Databricks, Columbia University, and UC Berkeley. The program brings together researchers and builders working on agentic data systems, infrastructure, caching, programmability, and AI agents at scale. 🗓️ May 26, 2026 Learn more: https://bauplanlabs.github.io/SAO-workshop/

• View post

  Stack Overflow helped developers stop solving problems in isolation. Introducing cq exchange, where agents can now share and retrieve experience-driven knowledge across private namespaces and a shared public commons. Works across Claude Code, Cursor, Windsurf, OpenCode, the browser, and the CLI. Try cq exchange: https://link.mozilla.ai/cq-exchange

• View post

  You shouldn’t need 15 tabs open to finish one task. Join us as Caroline Bohu, Solutions Engineer at https://mozilla.ai, walks through how AI agents can help automate repetitive workflows across your existing tools using Octonous. We’ll cover practical workflow examples, connected tools, approvals, recurring automations, and how Octonous can help reduce manual work across ops, support, sales, product, and marketing teams. 📆 June 8 at 2 PM BST 📍 Online Register here: https://link.mozilla.ai/oc...

• View post

  If you missed John Dickerson’s piece on sovereign AI a couple weeks ago, one section feels increasingly relevant: Many AI systems today depend heavily on one provider’s pricing, policies, and access remaining stable over time. The piece looks at how teams can build with more resilience and flexibility across their stack. Read more: https://link.mozilla.ai/sovereign-ai

• View post

  If you missed John Dickerson’s piece on sovereign AI a couple weeks ago, one section feels increasingly relevant: Many AI systems today depend heavily on one provider’s pricing, policies, and access remaining stable over time. The piece looks at how teams can build with more resilience and flexibility across their stack. Read more: https://link.mozilla.ai/sovereign-ai

• View post

  AI is useful. Until you have to connect it to real work. On June 8 at 2PM BST, Caroline Bohu from @MozillaAI will host a live Octonous webinar focused on practical AI workflows for everyday work. We’ll walk through connected workflows, approvals, automations, and practical examples across ops, support, sales, product, and marketing teams. Join the live walkthrough: https://link.mozilla.ai/octonous-june2026-webinar

• View post

  The shift back to local AI isn’t only about privacy anymore. For a lot of teams, it’s becoming a cost decision. In this new post, Anushri Gupta, DevRel at @MozillaAI, looks at rising cloud AI costs, local AI lock-in, and why portability is becoming a bigger conversation for developers. Read it here: https://link.mozilla.ai/ai-got-expensive

• View post

  What happens when you turn a local AI model into a coding agent? We tested Gemma 4 26B running locally through llamafile on a MacBook GPU. The model connected to VS Code through an OpenAI-compatible server. Then we gave it a broken React app to fix. It read the codebase, rewrote components, and fixed the cart issue completely offline. Try it yourself: https://link.mozilla.ai/huggingface-llamafile-010

• View post

  Open-weights models should not feel like a downgrade. Today we're launching Otari, an open-source LLM gateway built on top of any-llm, and https://otari.ai, the hosted platform built around it. Use frontier or open-weights models without giving up web search, code execution, observability, budget controls, routing policies, and more. Try the beta: https://link.mozilla.ai/otari-own-your-ai-stack

• View post

  Meet Carlos Álvarez Barba, who recently joined https://mozilla.ai. Carlos has built and operated infrastructure throughout his career, from high-performance computing to cloud, platform engineering, data platforms and AI/ML systems. He enjoys solving the behind-the-scenes challenges that keep systems reliable, deployments smooth and developers productive. Outside of work, he’s passionate about tropical plants, paludariums and side projects. Say hello and welcome Carlos to the team 👋

• View post

  Less than one week until our live Octonous webinar. A lot of team productivity gets lost between tools, handoffs, approvals, and manual follow-ups. Join Caroline Bohu from @MozillaAI for a practical walkthrough of how AI agents can help teams automate repetitive workflows and reduce busywork. Save your spot: https://link.mozilla.ai/octonous-june2026-webinar

• View post

  Can healthcare organizations build better AI without sharing sensitive patient data? On June 10 at SIIM26, John Dickerson, CEO of https://mozilla.ai, will present "Share For Healthcare: Privacy-Preserving, Open Access to Data and Models Is Critical in the Age of AI." The session will explore how institutions can collaborate on AI development, improve models, and keep patient and provider data local and secure. Get the details: https://link.mozilla.ai/siim26

• View post

  What is an LLM Control Plane, and why do AI teams need one? Anushri Gupta (Founder, DevRel at https://mozilla.ai) explains how control planes improve reliability & governance for LLMs. Read: https://blog.mozilla.ai/what-is-an-llm-control-plane/ #AI #LLM

• View post

  Image classification is now supported in encoderfile 📸 encoderfile makes it easy to deploy encoder models with zero dependencies. With this release, we're expanding beyond text tasks to support image classification, including flexible Lua-based preprocessing and support for CLI, HTTP/S, and gRPC interfaces. Coming next: - Object detection and image segmentation 👀 Read the full post and let us know what features you'd like to see next: https://blog.mozilla.ai/image-classificat...

• View post

  Product managers spend too much time moving between Slack, GitHub, Notion, and countless other tools. In our latest blog post, we share how we use Octonous inside @MozillaAI to reduce that overhead. Some of the workflows we rely on every day include capturing feedback, generating product-focused PR summaries, and searching across connected tools, all while keeping humans in control. 👀 Read the full article: https://blog.mozilla.ai/using-octonous-as-a-product-manager/

• View post

  Introducing transcribe.cpp: a new open-source C/C++ speech-to-text inference library developed through https://mozilla.ai's Builders in Residence program. Built on ggml, transcribe.cpp provides: 🎙️ Support for multiple STT model families via GGUF ⚡ GPU acceleration with Metal, Vulkan, and CUDA 💻 Portable, local inference across platforms 🔓 An open foundation for building fast transcription into your own applications. Built through https://mozilla.ai's BiR program. 👀 Read the ar...

• View post

  What is any-llm? One interface to talk to any model. Across local and self-hosted setups. Missed the livestream? Watch the recording: https://link.mozilla.ai/osf-any-llm

• View post

  Meet CJ, one of our newest Builders in Residence. He’s contributed to llamafile, created whisperfile, and built Handy. At Mozilla.ai, he’s focused on exploring new ways to distribute and run local AI.

• View post

  Agent memory is the new moat and it needs to be open. Earlier last month we wrapped two days at the Global Portable Memory Workshop in Seattle. The room was full of people who get it: if portable, user-controlled memory consolidates inside closed platforms, users lose. We need open infrastructure for the AI era, the same way Linux and open protocols built the web. Grateful to Gates Foundation, @mozilla, Anthropic, and the full crew. #PortableMemory #AgenticAI #OpenSource

• View post

  Agents run into the same issues over and over. 1k GitHub stars in, that’s still the pattern we see. cq is shared memory so they don’t have to figure things out from scratch each time. Open source, still early: https://link.mozilla.ai/cq-repo

• View post

  Running a trade business means more than just doing the job. Nathan Brake, Senior ML Engineer at https://Mozilla.ai, was featured on Pittsburgh Tech Beat. The article highlights the admin work that fills nights and weekends, and how Clawbolt is being built to help manage it. https://www.pghtechbeat.com/p/an-ai-assistant-for-the-trades

• View post

  We’re ecstatic to welcome @totostache to https://Mozilla.ai! He’s a high-energy software engineer who loves coffee, photography, and open-source. He’s joining the team working on the any-llm managed platform, helping teams work more smoothly across different LLM providers.

• View post

  @fosdem 2026 kicks off this weekend (Jan 31–Feb 1) in Brussels. If you’re attending, come visit the Mozilla booth and say hello. @totostache, Nathan Brake, and @mala will be there throughout the event. See you at FOSDEM!

• View post

  Encoder models don’t chat, but they power most production NLP. Deploying one still means dragging along a Python runtime, a full dependency tree, and serving infrastructure designed for something much larger. Encoderfile takes a different approach. A pre-built executable with an appended payload containing model weights, tokenizer data, and a manifest. At runtime, it reads itself and loads everything accordingly. Full post: https://link.mozilla.ai/encoderfile-new-format

• View post

  https://Mozilla.ai is now a launch partner of Flower Hub. We published a new project called **fed-phish-guard**, which trains a phishing URL classifier using federated learning. Each client trains locally and only sends model updates back to the server, keeping browsing data private. Read the blog post here: https://link.mozilla.ai/flower-hub Check the project page here: https://flower.ai/apps/mozilla-ai/fed-phish-guard/

• View post

  Daniel Nissani (AI Safety Engineer) and Roya Pakzad (Senior Mozilla Fellow) evaluated three guardrails on multilingual humanitarian prompts: • FlowJudge • Glider • AnyLLM (GPT-5-nano) Each reviewed the same asylum-related scenarios in English and Farsi. Details: https://link.mozilla.ai/multilingual-guardrails