blog: hivemind goes AGPL (#3)

* blog: hivemind goes AGPL

* blog: hivemind goes AGPL

* blog: hivemind goes AGPL

* blog: hivemind goes AGPL

* Update public/blogs/2026-01-13-hivemind_agpl.md

Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>

* blog: hivemind goes AGPL

---------

Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>

Author: JarbasAl

public/blogs/2026-01-13-hivemind_agpl.md

@@ -0,0 +1,90 @@
+---
+title: "HiveMind-core v4.0: AGPL, History, and the Path Forward"
+excerpt: "For years, I’ve been chasing business leads and grant opportunities just to keep the lights on, which means HiveMind ideas often stayed in my head; some lost forever, some only half-remembered. This is not sustainable, and it’s not fair to the project or its potential users."
+coverImage: "/assets/blog/hivemind_agpl/thumb.png"
+date: "2026-01-13T00:00:00.000Z"
+author:
+  name: Casimiro Ferreira
+  picture: "https://avatars.githubusercontent.com/u/33701864"
+ogImage:
+  url: "/assets/blog/hivemind_agpl/thumb.png"
+---
+
+
+# HiveMind-core v4.0: AGPL, History, and the Path Forward
+
+Sometimes necessity doesn’t just spark invention, it drags you along and forces you to make something you didn’t know you needed. HiveMind started exactly like that.
+
+Around **2015**, I moved from Windows to Linux. Soon after my main laptop broke, leaving me stuck with a barely functional **32-bit relic**. Mycroft was too heavy to run locally, but a friend lent me access to a server (hello Chris Schantz, if you’re reading this!). The plan: run Mycroft on the server and access it from my laptop. To do that, I needed middleware to bridge the two. And thus, **HiveMind** was born — a tiny, chaotic, stubborn little project that would grow into a full network-first voice middleware.
+
+This early hack pushed me deeper into Mycroft contributions and eventually led to the creation of **OpenVoiceOS (OVOS)**. But let’s be clear: **HiveMind is my personal project, not an OVOS Foundation project**. 
+
+OVOS integration happens naturally; it’s modular, and HiveMind can leverage its plugins: STT, TTS, WakeWord, and even the listener code for satellites. But you don’t need OVOS at all, HiveMind can run **any “brain” or agent**, from OVOS to a straight-up LLM. Satellites don’t care; they just work. Modularity rocks.
+
+---
+
+## HiveMind Through the Ages
+
+HiveMind has evolved through necessity, experimentation, and stubbornness:
+
+1. **TCP sockets (2015)** - proof-of-concept: can two machines even talk?
+2. **WebSockets** - more reliable communication, fewer broken dreams.
+3. **Multiple simultaneous transport protocols** - HTTP on one port, WebSockets on another, satellites don’t blink.
+4. **Modular plugin architecture** - custom transports, binary audio streams, fully replaceable brains. Drop-in compatible, edge devices oblivious.
+
+Now HiveMind is a **network-first transport + authentication middleware**, designed for distributing agents (brains) across any number of devices. Replace the brain? Just plug in a different agent. Use an LLM? Satellites don’t care. Permissions may not fully apply in that case (no skills/intents to allow/block), but all infrastructure still works.
+
+---
+
+## License Change: AGPL as a Necessary Evil
+
+Let’s be honest: AGPL is not glamorous. I’ve criticized GPL before; the “my kind of freedom is better than yours” discourse gets old. GPL projects can use my Apache code, but I can’t use theirs. HiveMind will become **less free for commercial usage**, but users arguably gain more protections. FSF and the Linux world would probably approve.
+
+Why AGPL?
+
+* **Networked deployments:** HiveMind is almost always used across multiple devices. AGPL ensures modifications remain open.
+* **Sustainability:** Commercial users who want to avoid AGPL will pursue a **commercial license**, letting me focus on HiveMind development instead of chasing business leads. Indirectly, OVOS also benefits, since it’s the reference agent implementation.
+* **FOSS sanity:** SSPL, open-core, or inventing a new license would be **FOSS suicide** or contribute to license proliferation. AGPL is the pragmatic compromise.
+* **GAFAM? Not our audience:** Big players will implement their own alternatives anyway. Copyleft doesn’t scare them into licensing HiveMind; and that’s fine. AGPL protects the community and small-to-medium businesses, not the already-resourced giants.
+
+> 💡 **Non-profits and permissively licensed projects** (MIT, BSD, Apache-2.0, etc.) are eligible for **no-cost commercial licenses**. Contributions to the ecosystem are considered sufficient — just reach out.
+
+---
+
+## Sustainable Development: Why This Matters
+
+Historically, HiveMind has often taken a backseat to OVOS. OVOS has a vibrant, active community, many eyes on the code, people willing to help with support, and lots of real-world testing. HiveMind, by contrast, is **niche, networked, and complex**. If I don’t work on it, it doesn’t get done. If I don’t provide support, no one else can.
+
+For years, I’ve been chasing business leads and grant opportunities just to keep the lights on, which means HiveMind ideas often stayed in my head; some lost forever, some only half-remembered. This is not sustainable, and it’s not fair to the project or its potential users.
+
+The AGPL change in v4.0 and the introduction of commercial licensing are **not about locking things down**. They're about creating a **reliable, funded pathway** to dedicate time to HiveMind development:
+
+* To implement long-standing features that have been waiting for years.
+* To improve stability, performance, and usability in real-world networked deployments.
+* To ensure that when someone asks “how do I do X with HiveMind?” there’s a real, maintained answer.
+
+In short, sustainable funding ensures HiveMind can grow **intentionally**, instead of being a series of half-baked ideas lost to time or postponed indefinitely. It’s about giving this project the attention it deserves and keeping the chaos under control, in the best possible way.
+
+---
+
+## Looking Ahead: OVOS 2026 and Beyond
+
+Thanks to the [**NLnet NGI Zero fund**](https://blog.openvoiceos.org/posts/2025-10-20-ngi), the **OVOS bus messages** will be standardized in 2026. Think **“OVOS Protocol via HiveMind”** — because apparently [everything is called a protocol in the age of LLMs, so why not us?](https://blog.openvoiceos.org/posts/2025-10-24-protocol_interoperability)
+
+* HiveMind used to be basically “get message from A to B.”
+* OVOS 2026 introduces a **defined inventory of valid messages**. Arbitrary messages are still possible.
+* Replacing the brain is standardized: a **hivemind-agent-plugin** translates OVOS Protocol messages to whatever your brain expects — OVOS, LLMs, A2A, Wyoming, whatever.
+
+This keeps HiveMind satellites fully **drop-in compatible**, modular, and future-proof.
+
+
+---
+
+## What This Means
+
+* HiveMind stays **open-source and usable**, even with AGPL.
+* OVOS continues as the reference agent, while HiveMind grows independently.
+* Commercial users now have a **clear licensing path**, funding sustainable development.
+* Sustainable, funded development, so I can finally tackle some of those long-forgotten ideas.
+
+From a broken laptop in university to fully-featured, network-first middleware, HiveMind has always been about **solving real problems, enabling communities, and letting chaos work beautifully**. AGPL is a **necessary evil**, ensuring HiveMind can grow sustainably, protect contributions, and avoid the trap of pseudo-FOSS licenses.

src/lib/blogUtils.ts

@@ -103,6 +103,7 @@ export const loadBlogPosts = async (): Promise<BlogPostMetadata[]> => {
       '2025-11-26-OVOS-hivemind-industry.md',
       '2025-12-09-ast.md',
       '2025-12-12-barranquenho.md',
+      '2026-01-13-hivemind_agpl.md',
     ];
 
     const posts: BlogPostMetadata[] = [];