# CQ Exchange: Shared AI Agent Knowledge _Mozilla.ai launches CQ Exchange, a centralized platform for AI agents to share and access experience-driven knowledge, moving beyond local storage._ **Updated:** 2026-08-22 **Published:** 2026-05-21 **Source:** https://www.startuphub.ai/ai-news/technology/2026/cq-exchange-shared-ai-agent-knowledge --- Mozilla.ai is releasing [CQ Exchange](https://blog.mozilla.ai/cq-exchange-agents-without-borders/), a new service designed to centralize and share knowledge among AI agents. This move follows positive community reception to the concept of 'Stack Overflow for Agents,' initially proposed in March. Siloed AI KnowledgeDriverFrom the article 9 mentionsPreviously, AI agent knowledge was siloed, either on local machines or self-hosted servers.problem leads toCQ Exchange LaunchCoreMozilla.ai centralizes AI agent experience-driven knowledge sharingFrom the article 5 mentionsMozilla.ai is releasing CQ Exchange, a new service designed to centralize and share knowledge among AI agents.introducesShared Knowledge StoreContextPlatform for agents to access and contribute knowledge unitsFrom the article 3 mentionsCQ Exchange, hosted by Mozilla.ai, offers a solution by providing a shared knowledge store.Global CommonsContextPublic repository for all agents to access curated KUsFrom the article 2 mentionsA key feature is the 'Global Commons,' a public repository for all agents.Private NamespacesContextFrom the article 3 mentionsAgents can now access their own private Knowledge Units (KUs) stored in private namespaces, accessible from any location via GitHub or Google authentication.Enhanced AI AgentsEffectAgents benefit from collective, experience-driven knowledgeFrom the article 9 mentionsThis move follows positive community reception to the concept of 'Stack Overflow for Agents,' initially proposed in March.Open-Source PrinciplesContextFrom the articleThis initiative emphasizes CQ's commitment to open-source principles, ensuring the command-line interface (CLI) and API remain consistent whether using CQ Exchange or a local instance.Community ContributionEffectFuture pipeline for community to contribute to the commonsFrom the article 3 mentionsMozilla.ai is seeding this commons with curated KUs, with plans to implement a community contribution pipeline soon. Previously, AI agent knowledge was siloed, either on local machines or self-hosted servers. CQ Exchange, hosted by Mozilla.ai, offers a solution by providing a shared knowledge store. Agents can now access their own private Knowledge Units (KUs) stored in private namespaces, accessible from any location via GitHub or Google authentication. ## The Public Commons A key feature is the 'Global Commons,' a public repository for all agents. Mozilla.ai is seeding this commons with curated KUs, with plans to implement a community contribution pipeline soon. This initiative emphasizes CQ's commitment to open-source principles, ensuring the command-line interface (CLI) and API remain consistent whether using CQ Exchange or a local instance. ## Accessing CQ Exchange CQ Exchange can be accessed through three interfaces: a web-based browser for human review and management, plugins and skills for AI agents (including Claude Code, OpenCode, Cursor, and Windsurf), and a CLI for terminal-based interactions. The browser interface allows users to review proposed KUs, manage API keys, and oversee the KU lifecycle. For agents, the plugin/skill integration allows querying CQ Exchange, proposing new KUs, and leveraging both private and public knowledge stores. The CLI offers similar functionalities, excluding KU reviews. ## Future Developments Upcoming features include a robust 'Graduation Pipeline' for community KU nominations and reviews, and 'Org namespaces' for secure, private knowledge sharing within organizations. This evolution shifts the focus from interfaces to the underlying data and knowledge agents utilize, potentially reshaping how applications are built around AI capabilities. It builds upon the foundational work of enabling AI agent knowledge sharing, aiming to prevent agents from repeating mistakes and improving efficiency. The project also incorporates early feedback loops, such as integrating VIBE✓ to help review knowledge units before they enter shared memory, as seen in efforts to enhance [Mozilla.ai cq](/ai-news/technology/2026/vibe-adds-friction-to-ai-coding-agents). --- Original analysis from [startuphub.ai](https://www.startuphub.ai), the #1 AI startup directory. © StartupHub.ai. All rights reserved. You may not republish this article in full without a license. Search engines and AI research tools may crawl and summarize for reference. Bulk reproduction or model training on this content requires a license. See https://www.startuphub.ai/terms.