# TechMaju AI Engine | Governed Enterprise AI Agents, On-Premises Ready

> Build enterprise AI agents with your data, tools and permissions. Works with leading models or local AI inference. On-premises, in Bahasa Indonesia and English.

Canonical page: https://techmaju.com/ai-engine/

## Summary

TechMaju AI Engine is a governed platform for building enterprise AI agents with your company's context, tools and permissions. It works with any leading model or your own local models.

## Product facts

| Fact | Detail |
| --- | --- |
| What it is | A platform for building governed enterprise AI agents with your context, tools and permissions |
| Built for | Operations, IT, security and support teams in regulated enterprises |
| Models | Compatible with OpenAI, Anthropic Claude, xAI Grok, DeepSeek, Z.ai and Qwen, or local inference clusters |
| Controls | Permissions, execution modes such as read-only, pre-run limits and a full trace of every action |
| Knowledge | TechMaju Knowledge Fabric, or your own documents and data |
| Work types | From single questions to multi-step agentic workflows |
| Languages | Bahasa Indonesia and English |
| Deployment | On-premises in your data center, or managed cloud hosted in Indonesia |
| Onboarding | Onboarding, implementation and enablement by TechMaju |
| Pricing | Enterprise license, quoted after a discovery session |

## A configured specialist

- Context: business data, documents and instructions keep each agent grounded.
- Skills: tools and workflows configured for the work each role does.
- Control: permissions, execution modes and pre-run limits define how far each agent can go.

## Models, governance and work

- One model gateway for text, vision, image and video models; providers can be swapped centrally.
- Compatible with models from OpenAI, Anthropic (Claude), xAI (Grok), DeepSeek, Z.ai and Qwen, and with local inference clusters.
- Every interaction is wrapped in permissions and execution modes, with full traces of conversations, outputs and usage.
- Agents handle single questions and multi-step agentic workflows, planning, acting and verifying across tools.

## Example patterns

- Help desk research: researches ticket history to help L2 staff diagnose issues; internal only, never replies to customers directly.
- Security investigation support: helps blue teams investigate endpoints and draft discovery reports in read-only mode.

## Frequently asked questions

### How is AI Engine different from a general-purpose AI chat assistant?

A general assistant works without your context or controls. AI Engine agents are configured with your data, tools and permissions, run in modes such as read-only, keep a full trace of every action, and can use local models on-premises.

### Is AI Engine tied to one model provider?

No. It is model agnostic, so you can switch or mix providers centrally without changing any agent's workflow.

### What is TechMaju AI Engine?

A platform for building governed enterprise AI agents with your company's context, tools and permissions, using any leading model or your own local models.

### Which AI models does it support?

It is model agnostic and compatible with models from OpenAI, Anthropic (Claude), xAI (Grok), DeepSeek, Z.ai and Qwen, configured from one gateway.

### Can it use local models on-premises?

Yes. It connects to sovereign local inference clusters, so model processing stays in your environment.

### Does it support Bahasa Indonesia?

Yes. Agents chat in Bahasa Indonesia and English.

### How do you control what an agent can do?

Each agent has permissions and an execution mode, such as read-only, set before it runs, and every action is traced.

### Can it use our company knowledge?

Yes. Pair it with TechMaju Knowledge Fabric, or ground agents in your own documents and data.

## Contact

Book a discovery session: hello@techmaju.com

PT TechMaju Digital Archipelago, Indonesia. https://techmaju.com/