Zulip vs Mattermost
Side-by-side trajectory, velocity, and editorial themes.
Zulip donates itself to a nonprofit foundation as its founder joins Anthropic.
Two consequential moves landed inside three weeks. Tim Abbott, Zulip's longtime leader, is stepping back to join Anthropic along with three senior Kandra Labs engineers, and the for-profit company is being donated to a newly formed independent Zulip Foundation. In parallel, Zulip Server 12.0 shipped — roughly 5,500 commits, end-to-end encryption for mobile push notifications, a major Docker upgrade, and configurable image previews — alongside routine 11.x security patches and a self-hosted AI search integration via Atolio.
Governance is being separated from product velocity: the foundation owns long-term stewardship while the technical roadmap (E2E push, self-hostable AI integrations, hardened install path) keeps targeting security-conscious, self-hosted teams. Losing the founder to a frontier AI lab is the kind of transition that either accelerates community ownership or stalls momentum — the 12.0 commit volume suggests the team built up runway before the announcement.
Expect the Zulip Foundation to publish formal governance, a maintainer charter, and a funding model in the next quarter, and for the AI integration story (Atolio-style connectors rather than first-party AI) to harden into Zulip's positioning against Slack and Mattermost's first-party AI bets.
v11.7 ships rearchitected AI agents and granular ABAC as Mattermost leans hard into regulated buyers.
Mattermost is now openly positioning as a collaboration platform for defense, intelligence, and critical infrastructure rather than a general-purpose team-chat alternative. The v11.7 release pairs Attribute-Based Access Control for Team Admins with a rearchitected Agents v2.0 layer that supports custom AI prompts and user-created agents, signaling that the AI roadmap will run on top of strict access governance rather than alongside it. Editorial output in May is overwhelmingly about sovereignty, coalition operations, and AI governance — the company is telling regulated buyers what to ask vendors during procurement.
The product is bifurcating from horizontal team chat into a sovereignty-and-governance-first platform aimed at procurement evaluations in defense and regulated finance. Each major release now ships more granular control surfaces (ABAC, coordinated ESR security cadence) underneath user-facing features (AI agents, custom prompts), which is consistent with a market where features only matter if they can pass a compliance review. Expect future releases to keep coupling AI capability to governance primitives rather than shipping AI features on their own.
The next minor release likely extends ABAC scope beyond Team Admins (channel-level or integration-level enforcement) and tightens the audit trail around user-created agents, since both are the natural follow-ons for a customer base that procures on control granularity. A coalition or cross-domain feature announcement is also plausible given how heavily April-May messaging leaned on multi-nation operational use cases.
See more alternatives to Zulip →
See more alternatives to Mattermost →