Zoho Corporation is expanding Catalyst, its cloud platform for building and deploying applications, with tools designed to let AI coding assistants perform more of the work involved in taking software from generated code to production.
The company said Monday that Catalyst by Zoho now supports Agent Skills, a non-interactive command-line interface and Model Context Protocol (MCP), alongside integrations with AI coding tools including Anthropic’s Claude Code and OpenAI’s Codex.
The move comes as AI coding assistants increasingly automate software development, shifting a growing part of the challenge from writing code to testing, deploying, managing and governing applications once they are built.
Zoho is positioning Catalyst as the infrastructure layer for that process. Rather than requiring developers to assemble separate services for functions, databases, hosting and AI capabilities, the platform provides those components within a serverless environment.
“With AI coding assistants, applications can be created very quickly, but moving that code into production reliably remains a challenge,” said Veerakumar Natarajan, Country Head, Zoho Kenya. “With Catalyst, we are reducing that friction by providing an agent-ready, serverless full-stack platform that helps developers build, deploy and scale intelligent applications with greater control and less operational overhead.”
From AI-Generated Code to Deployment
AI coding tools have made it easier for developers to generate application code, but the resulting software still needs infrastructure, deployment workflows and operational controls before it can be used in production.
Catalyst Agent Skills are designed to give coding assistants information about Catalyst’s services, architecture and recommended development practices. Zoho says the skills help AI agents understand which Catalyst services to use and how to work with them.
The platform’s non-interactive CLI is designed to allow AI coding assistants to execute multiple Catalyst operations without requiring developers to provide manual input at every step.
MCP support takes the integration further by allowing compatible AI development environments to interact directly with Catalyst capabilities.
Developers can, for example, perform database operations and other platform tasks from their existing AI development environments rather than repeatedly switching between their coding environment and the Catalyst console. Zoho’s Catalyst MCP documentation explains how developers can connect compatible AI clients to Catalyst.
Zoho said orchestration within Agent Skills can route requests through the appropriate CLI or MCP workflow, reducing the possibility of an AI assistant selecting the wrong tool or process.
The company has also made the Agent Skills repository open source and said it is open-sourcing Catalyst’s SDK, CLI and Slyte JavaScript framework.
Human Oversight Remains Part of the Workflow
The increased role of AI agents in software development has also raised questions about how much autonomy should be given to automated systems, particularly when they have access to production infrastructure.
Zoho said Catalyst maintains a separation between development and production environments. Code must be promoted to production manually, preventing AI agents from directly accessing production environments.
The platform also provides scoped collaborator controls, allowing organizations to determine who or what can participate in development and which actions they can perform.
Audit trails record application activity, platform activity and MCP tool calls, while changes made after launch are versioned and can be attributed, tracked and reversed, according to Zoho.
Applications running on Catalyst are hosted in Zoho’s own data centers and use the company’s security infrastructure, including distributed denial-of-service protection, SOC compliance, vulnerability assessments, penetration testing and a web application firewall.
For software companies and system integrators, those controls could become increasingly important as AI agents move beyond code generation and begin executing operational tasks on behalf of developers.
Free Access for Students
Alongside the new agentic development capabilities, Zoho is removing the cost barrier for students using Catalyst to build non-commercial applications.
Students can deploy applications without a subscription or credit card and receive access to Catalyst’s full-stack development capabilities, including hosting, functions, databases and AI tools.
Catalyst’s student program is aimed at giving students practical experience building and deploying complete applications rather than limiting their exposure to writing code.
Zoho says students can use Catalyst for qualifying non-commercial projects at no cost, with up to $250 in usage over six months. Students are not automatically charged if they reach that threshold; they can request additional free usage if their projects qualify.
For other developers, Catalyst offers a monthly free tier and a pay-as-you-go pricing model, with individual features charged according to usage rather than through an additional software license.
Developers can monitor usage through the Catalyst console and set budget alerts and spending ceilings. A subscription option is also available for teams seeking more predictable costs.
Catalyst is available for immediate use.
The broader strategy reflects a shift in cloud computing as AI coding tools reduce the time required to produce software. As code generation becomes increasingly commoditized, cloud providers and platform companies are competing to control the infrastructure, deployment, security and governance layers that determine how that software reaches users.
For Zoho, Catalyst’s integration with AI coding assistants is an attempt to position its cloud platform within that emerging workflow, while retaining human approval over the final transition to production.