Ho Chi Minh City · Since 2024

Software that works.
AI that earns its place.

We build AI-first products where you just ask, and the software does the rest. Powered by large language models, with agentic workflows, MCP servers, Skills and plugins. We also build custom software for businesses.

  • AI-first apps
  • AI agents
  • RAG
  • MCP servers
  • Skills & plugins
  • Web & mobile
  • Cloud & DevOps
Expense trackerConcept preview

coffee 45k, Grab home 80k

Logged 2 expenses:
  • Food & drink45,000₫
  • Transport80,000₫
Food spending is on track for this month’s budget.

where can I cut back?

MCP · add_expense MCP · query_spending Skill · monthly review

Products · in development

AI-first products, built on LLMs

Each product puts an AI assistant at the centre. You ask in plain Vietnamese or English instead of clicking through forms and menus.

Consumer · first product

AI-first personal expense tracker

Say what you spent, like coffee 45k, Grab home 80k, and the AI records and categorises it. Ask how much did I spend on food this month? or where can I cut back? and get a straight answer.

Skills cover monthly reviews, budget alerts and receipt capture.

Business · next

AI-first products for business

Agentic workflows that take a goal, plan the steps and carry them out across your existing systems, with a person approving what matters.

Plugins and MCP connectors link agents to the tools a company already uses. Skills capture each team’s procedures, so every run is consistent, auditable and easy to review.

MCP servers

Each product exposes its data and actions through the Model Context Protocol, with scoped permissions, so AI assistants can work with it securely.

Skills

Packaged, tested workflows for recurring jobs, so results are consistent rather than improvised.

Plugins

MCP servers and Skills bundled together, so users install a whole product into their assistant in one step.

  1. ResearchUse cases, model evaluation, architecture
  2. BuildExpense tracker, MCP server and Skills
  3. PilotEarly access with selected users
  4. ExpandAgentic AI products for business

AI engineering

Agents that do the work, grounded in your data

The three building blocks that make AI useful inside a business: agents that act, knowledge they can trust, and secure connections to the systems you already run.

  1. Goal“Reconcile this month’s supplier invoices”
  2. PlanThe agent breaks the goal into steps
  3. RetrieveRAG finds the contracts and records it needs
  4. ActMCP tools read and update your systems
  5. ReviewA person approves what matters
  6. DoneResult, sources and a full log

AI agents & workflows

Agents that take a goal, plan the steps and use tools to finish the job, from triaging requests to preparing reports. People approve the steps that matter, and every run is logged and evaluated.

  • Planning
  • Tool use
  • Human-in-the-loop
  • Evaluation

RAG: answers from your own data

Retrieval-augmented generation grounds the AI in your documents, policies and databases, with citations, so it says what your data says instead of guessing. Access rules decide who can see what.

  • Vector & hybrid search
  • Citations
  • Access control
  • Freshness

MCP: connect AI to your systems

Model Context Protocol servers give AI assistants a standard, secure way to read and act on your CRM, ERP, databases and internal APIs, with scoped permissions and audit trails.

  • MCP servers
  • Scoped permissions
  • Audit logs
  • Skills & plugins

IT services

From first prototype to production

One team that owns the outcome, for companies that need software built well.

Custom software

Web applications, internal tools and APIs designed around how your business actually runs.

AI & LLM integration

Assistants, document search (RAG), MCP servers and workflow automation wired into your existing systems.

Web & mobile apps

Fast, accessible front-ends and cross-platform mobile apps, built to ship and easy to maintain.

Cloud & DevOps

Infrastructure, CI/CD, monitoring and cost control on AWS, GCP, Azure and Cloudflare.

Data engineering

Pipelines, warehouses and dashboards that turn scattered data into something you can decide with.

IT consulting

Architecture reviews, technology selection and hands-on support for teams that need a second opinion.

How we work

Small steps, visible progress

  1. 01

    Discover

    We learn the problem, the users and the constraints before writing code.

  2. 02

    Design

    A clear plan, scope and architecture you can review and challenge.

  3. 03

    Build

    Short iterations with working software you can try every week.

  4. 04

    Operate

    Launch, monitor and improve. We stay on after go-live.

Have a project in mind?

Tell us what you're building, or ask to join early access. We usually reply within one business day.

[email protected]

Company details →