Label
GeneXus for Agents FAQ
What is GeneXus for Agents?
GeneXus for Agents is a set of skills, tools, and interfaces that enable AI agents to operate on GeneXus Knowledge Bases in a controlled way and with platform-specific context.It is a capability layer that makes it possible to use GeneXus from environments the market is already adopting, such as CLIs, coding assistants, and other agents compatible with skills, tools, and MCP.
In practice, GeneXus for Agents enables a new way of working: Knowledge Base objects can be represented as text, understood and modified by agents, validated against the KB, and reintegrated into the development workflow without relying exclusively on the IDE.
What is GeneXus for Agents used for?
GeneXus for Agents is designed to incorporate AI agents into workflows involving GeneXus Knowledge Bases, with platform-specific context and without losing the advantages and guarantees of GeneXus’s deterministic generation engine.It allows agents to understand, generate, review, and propose changes to GeneXus artifacts without depending on the IDE.
In practice, it helps accelerate analysis, development, maintenance, and software evolution by integrating GeneXus with the environments, agents, and workflows already used across the industry.
What are GeneXus Skills?
GeneXus Skills is the set of skills, references, conventions, and tools that allow AI agents and coding CLIs to understand, generate, review, and modify GeneXus artifacts with platform-specific context.In other words, if GeneXus for Agents is the overall offering, GeneXus Skills is its specialization layer. It’s what enables that work to happen in a useful and consistent way, both within recommended environments and in other CLIs or market agents that support this model, such as Claude Code, Globant CODA, or OpenAI Codex.
GeneXus Skills provides operational context. It encapsulates best practices, structure, references, and rules so agents can work more effectively with GeneXus.
This allows agents and coding CLIs to operate on Knowledge Bases and GeneXus artifacts with greater consistency, speed, and accuracy.
What benefits does GeneXus for Agents provide?
Expanded productivity
Unlike the traditional model - where much of the work happens sequentially and is centered around direct interaction with the IDE - with GeneXus for Agents, developers can rely on agents to explore, generate, review, and propose changes across different modules or tasks, with greater speed and less friction.
This changes how work is done. Instead of manually modifying knowledge bases, the focus shifts to coordinating and supervising agent-assisted work.
GeneXus continues to provide something unique: generating and maintaining mission-critical systems for decades, at scale, with continuous technological evolution, in a deterministic and correct-by-construction way, at a fraction of the cost of generative AI. These strengths remain unchanged.
What GeneXus for Agents adds is the ability to combine that core with what agents do best: exploring knowledge, generating frontends, producing auxiliary code, creating integrations, accelerating repetitive tasks, and assisting in system evolution.
GeneXus handles the critical business core. Agents expand operational capacity around that core. Together, they allow teams to cover more work faster while making better use of existing knowledge.
It also represents early steps toward adopting:
Importantly, the GeneXus engine does not change. Again: code generation remains deterministic and correct by construction. Semantic validations, normalization, impact analysis, and more are always applied. What changes is how the Knowledge Base is accessed and operated on.
GeneXus for Agents increases productivity, expands what a team can accomplish with the help of agents, and opens GeneXus to modern development workflows - without sacrificing control, consistency, or determinism. In one sentence:
GeneXus for Agents turns the Knowledge Base into a real workspace for agents, without losing the guarantees of the GeneXus engine.
By working with textual representations of the Knowledge Base and UX requirements within the same workflow, agents can help build interfaces using external technologies such as React or other modern frameworks - without disconnecting those interfaces from the business knowledge and rules defined in GeneXus.
This enables a particularly valuable approach for many teams: using GeneXus to generate and maintain the system’s core - with all its robustness, consistency, and evolution capabilities - while leveraging other technologies in the frontend when greater flexibility, dynamism, or design freedom is needed.
In other words, GeneXus for Agents allows teams to combine the best of both worlds: a solid, deterministic backend powered by the Knowledge Base, and a more creative, modern user experience layer.
This changes how work is done. Instead of manually modifying knowledge bases, the focus shifts to coordinating and supervising agent-assisted work.
Expanded capabilities without losing GeneXus’s core
GeneXus for Agents complements GeneXus’s traditional deterministic generation - secure, reliable, maintainable, and low-cost - with the new capabilities brought by LLMs and AI agents.GeneXus continues to provide something unique: generating and maintaining mission-critical systems for decades, at scale, with continuous technological evolution, in a deterministic and correct-by-construction way, at a fraction of the cost of generative AI. These strengths remain unchanged.
What GeneXus for Agents adds is the ability to combine that core with what agents do best: exploring knowledge, generating frontends, producing auxiliary code, creating integrations, accelerating repetitive tasks, and assisting in system evolution.
GeneXus handles the critical business core. Agents expand operational capacity around that core. Together, they allow teams to cover more work faster while making better use of existing knowledge.
Modern development workflow
GeneXus for Agents aims to integrate GeneXus with how the industry already works today:- Any compatible LLM or agent can act as a copilot, with no lock-in to a single provider
- MCP as an interoperability protocol to connect agents with GeneXus capabilities
It also represents early steps toward adopting:
- Git as the source of truth for versioning, branching, pull requests, and code review
- Standard CI/CD to automate build, validation, and deployment from pipelines
More openness, same control
Importantly, the GeneXus engine does not change. Again: code generation remains deterministic and correct by construction. Semantic validations, normalization, impact analysis, and more are always applied. What changes is how the Knowledge Base is accessed and operated on.
In summary
GeneXus for Agents increases productivity, expands what a team can accomplish with the help of agents, and opens GeneXus to modern development workflows - without sacrificing control, consistency, or determinism. In one sentence:GeneXus for Agents turns the Knowledge Base into a real workspace for agents, without losing the guarantees of the GeneXus engine.
UX in other technologies
GeneXus for Agents also expands possibilities in the user experience layer.By working with textual representations of the Knowledge Base and UX requirements within the same workflow, agents can help build interfaces using external technologies such as React or other modern frameworks - without disconnecting those interfaces from the business knowledge and rules defined in GeneXus.
This enables a particularly valuable approach for many teams: using GeneXus to generate and maintain the system’s core - with all its robustness, consistency, and evolution capabilities - while leveraging other technologies in the frontend when greater flexibility, dynamism, or design freedom is needed.
In other words, GeneXus for Agents allows teams to combine the best of both worlds: a solid, deterministic backend powered by the Knowledge Base, and a more creative, modern user experience layer.
Availability and licensing
Is there a beta testing option? When can I try it?
GeneXus for Agents has been available since March 31, 2026.How will GeneXus for Agents be licensed?
GeneXus for Agents and its skills will be available at no cost to all customers who are up to date with maintenance.How do I get started? Where can I find more information?
https://docs.genexus.com/en/wiki?61619,GeneXus+for+AgentsCompatibility and versions
Which versions of GeneXus are compatible with GeneXus for Agents?
GeneXus for Agents supports Knowledge Bases from GeneXus Next and GeneXus 18.- For GeneXus Next KBs, agents work directly on the base.
- For GeneXus 18 KBs, the workflow uses GeneXus Next to create a local checkout. Agents work on that local copy, and once changes are approved, the KB is republished to GeneXus Server for deterministic generation with GeneXus 18.
Can it be used with older versions like GeneXus 15, 16, or 17?
Not at the moment. To use GeneXus for Agents, the KB must be in a compatible version. If your KB is on an older version, the first step is to migrate it to GeneXus 18.So is the correct approach to move KBs to GeneXus Next and then use GeneXus for Agents?
Yes. From March 2026 until the next release, GeneXus for Agents works on GeneXus Next Knowledge Bases. Soon it will also support GeneXus 18 directly.How do I migrate from GeneXus 15 to GeneXus 18?
Globant’s Fast Code Studio and GeneXus partners provide tools and processes to help accelerate this migration path.Does this replace the IDE?
No. The GeneXus IDE - whether GeneXus Next or GeneXus 18 - remains a valid option. What changes is that it is no longer the only option.There are now three ways to interact with the Knowledge Base:
- Through the IDE
- Through coding agents (CLIs)
- Through any agent that supports Skills and the MCP protocol
LLMs, Skills, and MCP
Which LLMs work with GeneXus for Agents?
Any LLM that supports MCP (Model Context Protocol):- CODA CLI with models via Globant Enterprise AI (recommended for security and control)
- Claude Code (Anthropic)
- Codex (OpenAI)
There is no lock-in to any specific LLM. Your team chooses the model.
What changes and what does NOT change with GeneXus for Agents?
What changes: How you access the Knowledge Base. In addition to the IDE, it can now be accessed via terminal, from any LLM, and through standard protocols like MCP.What does NOT change: The GeneXus engine. The runtime is not modified, generators are not altered, and the core is not rewritten. Code generation remains deterministic and correct by construction.