---
title: syntax.md
description: syntax is an independent software studio. We design and build fast, agent-native web products with a focus on performance, architecture, and craft.
email: contact@syntax.md
lang: en
base: Chișinău
---

# Software built with care.

syntax is an independent web software studio. We design and build fast, agent-ready products with deep attention to craft, performance, and everyday usability.

[Start a project →](mailto:contact@syntax.md?subject=New%20project)

## Core principles.

Every product we ship is evaluated against the same clear standards.

### 01 — Agents

**AI that executes real actions.** Autonomous workflows that call tools, handle failures predictably, and maintain full auditability.

### 02 — Performance

**Speed is foundational.** We define strict performance budgets from the start, enforce them in CI, and treat latency regressions as critical bugs.

### 03 — Details

**Craft in the small details.** Transitions, error handling, edge cases, and keyboard navigation. The subtle polish that makes software feel solid.

## What we build.

A focused team delivering end to end. From architecture and design to production, without hand-off gaps.

1. **Web applications** — Scalable, high-performance web applications built with React, Next.js, and TanStack. Type-safe from the database layer to the user interface. `react` `next` `tanstack`
2. **AI agents & MCP servers** — Production-grade agents and MCP servers that connect your product to AI models, validated with automated benchmarks. `ai-sdk` `mcp` `opencode`
3. **Design engineering** — Polished interfaces and design systems translated faithfully from Figma into clean, accessible production code. `figma` `design-systems` `motion`
4. **Performance optimization** — Comprehensive audits and profiling. We eliminate bottlenecks, cut load times, and put automated guardrails in place to prevent regressions. `core-web-vitals` `profiling`
5. **Platform & infrastructure** — Modern monorepos, automated CI pipelines, and end-to-end testing suites. The foundational tooling that keeps your team moving fast. `turborepo` `effect` `e2e`

## Your product, agent-ready.

We design the tool interfaces, context layers, and guardrails that allow AI models to interact reliably with your stack.

- Tool & MCP server architecture
- Automated evaluation suites
- Durable workflows & observability
- Human oversight on critical actions

## Tools we rely on.

Pragmatic choices: battle-tested foundations for stability, modern tooling for speed and leverage.

### Languages & frameworks

| tool | role | why |
| --- | --- | --- |
| [React](https://react.dev) | `ui` | Server Components, optimistic UI, and modern concurrency. |
| [TanStack](https://tanstack.com) | `state` | Query, Router, Table, and Form. Predictable client state management. |
| [Effect TS](https://effect.website) | `runtime` | Type-safe errors, resilient retries, and structured concurrency. |
| [AI SDK](https://ai-sdk.dev) | `models` | Unified streaming, tool calling, and structured outputs across models. |
| [OpenCode](https://opencode.ai) | `agents` | Open-source agent infrastructure for code generation and automation. |

### Tools & infrastructure

| tool | role | why |
| --- | --- | --- |
| [Turborepo](https://turborepo.com) | `build` | High-performance monorepos with intelligent caching. |
| [e2e](https://tester.army/e2e) | `testing` | Resilient end-to-end testing suites integrated into CI. |
| [Figma](https://figma.com) | `design` | Design tokens and component libraries synced with code. |
| [MCP](https://modelcontextprotocol.io) | `protocol` | The open standard connecting AI models to external data and tools. |
| [Zed](https://zed.dev) | `editor` | High-performance, collaborative code editor built for speed. |

## How we work.

No unnecessary meetings or slide decks. Visible, regular progress, direct updates, and working software.

```sh
$ git log --oneline
a1f3c09  init     discovery call
7be21d4  spec     technical scope & roadmap
c90e8f2  build    continuous preview builds
e4d71aa  ship     production release
HEAD     iterate  refinement & handover
```

- **init** (start) — A focused discussion to understand your requirements, technical scope, and whether we're the right fit.
- **spec** (scope) — A clear specification outlining deliverables, scope, and how we'll work together.
- **build** (ongoing) — Regular working builds instead of repetitive status meetings. Direct reviews and continuous feedback.
- **ship** (launch) — Observability, testing, and documentation verified beforehand for a smooth, controlled release.
- **iterate** (post-launch) — We monitor stability post-launch, address edge cases, and hand over everything your team needs.

## Let’s talk.

Tell us what you’re building, your technical requirements, and your target timeline.
We'll get back to you soon.

→ [contact@syntax.md](mailto:contact@syntax.md)

---

[Terms](https://syntax.md/terms) · [Privacy](https://syntax.md/privacy) · [llms.txt](https://syntax.md/llms.txt)