Google Cloud launches Gemini Factory to speed up integrations

Google Cloud is establishing a "Gemini factory" to assist publishers in building and publishing integrations and agents on Google Cloud Marketplace. This initiative combines workshops, engineering tools, and distribution mechanisms, with governance and security options tailored to the needs of large enterprises. Several partners have reached production in just a few days.

Credits, Marketplace, and Integrated Discovery for Partners

Google Cloud offers initial credits to offset construction and porting costs, as well as performance credits based on consumption and targeted technical support. Partners can upload their solutions directly to Google Cloud Marketplace, enabling simplified management of billing, purchasing, and integration within enterprises, while providing them with international exposure to enterprise and mid-market clients. The offering also includes co-marketing resources, social media promotion activities, public relations kits, launch operations, and funded packages for top-performing partners. The deployment architecture aims to enable seamless client integration and ensure the protection of intellectual property, through access via native OAuth, Dynamic Client Registration, and simplified BYOL processes, while the partner retains control over backend security, data, and proprietary tools in their environment. Solutions are also highlighted in the AI Solution Finder, which centralizes the discovery of validated MCP connectors, plugins, and agents, facilitating configuration and governance by IT administrators. Google Cloud indicates that it supports verified partners at every stage, with startup credits at publication, performance credits for inference, and GTM and commercial resources aimed at transforming engineering milestones into a pipeline.

Four Elements to Integrate Third-Party Capabilities into Gemini

Previously, third-party integrations operated as isolated agents, forcing employees to step out of their workflow. The program encourages providers to think in four complementary elements. MCP connectors serve as a channel to expose external APIs, search capabilities, and CRUD operations, registered as callable tools in the Agent Registry. Declarative skills, open Markdown configurations stored in the Skills Registry, provide business instructions, operational guardrails, and repeatable multi-step workflows. Plugins, available in Gemini and on the Marketplace, package connectors and skills into turnkey solutions governable by IT and callable by users. For proprietary algorithm needs, compliance, or deterministic execution, A2A agents operate in an isolated partner environment and integrate as callable connectors, protecting intellectual property, ensuring deterministic sequences, and providing tamper-proof audit trails. These agents target ISVs with proprietary algorithms, regulated workflows, or strict compliance rules. Gemini unifies these components, shares session context with connectors, and chains outputs within the enterprise ecosystem. In this unified experience, companies can integrate third-party capabilities via MCP, A2A environments, REST APIs, declarative skills, and plugins, with the factory assisting in choosing the appropriate architecture.

An AI-Assisted Engineering Pipeline with Antigravity, RAD, and ADK

Once the architecture is set, months of integration are compressed, according to Google Cloud, into an AI-assisted pipeline led with partner teams. In Antigravity, an AI-focused development environment, teams use pre-built Rapid Agent Development skills and guided workflows to translate specifications into functional architectures, build remote MCP connectors, agent environments, write .md skills, and launch automated validations. The current path involves exposing APIs via an MCP connector hosted on Cloud Run or on the partner's cloud and bundling skills into a turnkey plugin. When closed-source models or strictly deterministic execution are required, the Google Agent Development Kit and the A2A protocol are employed to expose execution as a callable connector in the Agent Registry. On the experience side, Gemini avoids external portals through native MCP applications and guided components, broadcasting contextual visual canvases in Gemini and chaining outputs via Google Workspace and other connectors. The program aims for a transition from use case to production in just a few days and accelerates client activation with incentives and go-to-market support.

Global Workshops and a Co-Innovation Track to Market

Google Cloud has launched a Gemini factory presented as an intensive co-innovation initiative to help partners move from whiteboard to agent solutions listed on Google Cloud Marketplace in record time. Hands-on workshops are conducted with Google Cloud specialists alongside partner teams to envision, architect, and build solutions with Gemini. These workshops take place in locations such as San Francisco, New York, Sunnyvale, London, Paris, Tel Aviv, and Tokyo. This track aims for direct publication on the Marketplace and global visibility with enterprise and mid-market accounts.

Nine Featured Partners and Sector Use Cases

Recent cohorts from the Gemini factory, previously known as Agent Factory, have produced industry-specific agents. Nine partners are highlighted, having reportedly reached production in just a few days, according to Google Cloud. The Aily Decision Intelligence agent aggregates 360° indicators to provide predictive diagnostics and strategic recommendations for executive decision-makers, leveraging complex operational and financial data to anticipate trends, identify supply chain disruptions, and guide resource allocation via a mobile application. FlowX offers a document reconciliation agent for financial markets, optimizing the extraction of institutional documents, multi-party reconciliation, and completeness audits of loan files, using document intelligence and agent environments based on Kubernetes, with real-time checks on contracts, regulatory filings, and credit documents to detect inconsistencies and significantly reduce back-office processing times. Klarvant's Digital Governance Analyst agent conducts an external assessment of an organization's namespace, encompassing post-quantum readiness, certificate/TLS hygiene, email security, supply chain risk, and external exposure; it provides detailed tracking of post-quantum migration, links supplier events and dependencies to impact points, and delivers responses in understandable language.