AI engineering studio

We build the infrastructure layer for AI agents.

Archetech AI Labs designs and ships the foundations teams need after the AI transition: agent skills, MCP servers, governance tooling and complete AI applications. Built to be secure, auditable and ready for production.

Book a 15-min call See our open source work

Open-source skills listed on Agensi
Works with Claude Code, Cursor, Codex and ChatGPT
Founder-led, every project

What we build

Start with the outcome, or bring us one piece of it. No generic AI wrappers.

Core service

AI-Native Operations

AI enablement and automation for your people, teams and processes.

We find where your team loses time, automate it with AI agents, and train your people to run and extend it. You end up with a measurable productivity gain and a team that can keep improving it.

Map
Find where time and effort go across your team's workflows.
Automate
Build AI agents and workflows for the highest-value tasks first.
Enable
Train your people and set clear rules, so the gains last after we leave.

Agent skills and MCP servers

Custom SKILL.md skills and MCP servers that give Claude Code, Cursor, Codex and other agents real capabilities: your workflows, your tools, your data. Packaged, versioned and documented so your team can maintain them.

AI governance and security review

An audit of how your team and agents use AI. We find exposed secrets, PII, risky permissions and prompt-injection paths across your AI tool configs, then put rules in place to keep them out.

Complete AI application builds

End-to-end products, from architecture to deployment. Outcome-based: we agree what "done" looks like before we start, and we ship to that.

Fractional AI architect

A senior architect on your side for the decisions that are expensive to get wrong: which models, which agent framework, how to structure context and memory, and how to roll AI out across a team.

Not sure which fits? Book a call and we will tell you honestly, including when you don't need us.

Built in the open

Read our code before you read our pitch. Everything here is free, open source and in daily use.

Mycelium Network, live. Drag to turn it.

Mycelium Network

Free, self-hostable community edition

A network of small, specialist AI models ("spores") that run on hardware you already own. Narrow tasks like redaction or extraction run locally, cheaply, and without data leaving your premises.

View project

Veil

Free, self-hostable community edition

The first spore on Mycelium. A local gateway that strips personal data from prompts before they reach Claude or ChatGPT, then restores it in the response. Built for teams handling sensitive data under India's DPDP Act.

GitHub

Agentic Playbook

Claude plugin, SKILL.md

Every AI coding session picks up where the last one left off. Agentic Playbook keeps a living playbook of your project's decisions, conventions and progress, so agents stop starting from zero.

GitHub Agensi listing

Agentic Prompt

Claude plugin, SKILL.md

Turns a rough ask into a precise spec. Agentic Prompt interviews you with quick, standard or deep questions, then hands your agent an exact, structured brief instead of a guess.

GitHub Agensi listing

The same engineering we put into our open-source work goes into every client project.

What we're building next

AthenOps Early access

AI governance for teams that use AI every day.

Your team runs Claude Code, Cursor and ChatGPT with dozens of prompts, skills and MCP servers. Most companies have no view of what those tools can access. AthenOps scans your AI tool and MCP configurations, flags exposed secrets, PII, risky permissions and prompt-injection risks, and enforces your rules from one dashboard.

What v1 does

Scans Claude Code, Cursor and MCP configurations with a lightweight capture agent
Flags secrets, PII, risky permissions, prompt-injection risks and config drift
Applies your organization's rules and shows every finding in one web dashboard

On the roadmap, not yet available: agent registry, runtime monitoring, guardrails, kill switch and full audit trail.

Small first steps. No big commitments.

A 15-minute call

You tell us the problem. We tell you honestly whether we're the right fit, and what we'd do first.
A time-boxed pilot

Two to three weeks, one clear deliverable, agreed in writing before we start. You see real work before any long-term decision.
Build and ship

Weekly demos, shared progress, and code and docs you own from day one.
Ongoing partnership, if it works

Maintenance, new features or a retained architect. Only if the pilot earned it.

You own everything we build for you: code, docs and IP.

We're new. Here's why that works in your favor.

Archetech AI Labs is a new studio, and we're taking on a small number of founding partners. We're telling you upfront because trust starts with honesty. Being new means you get things larger agencies can't offer.

Apply to be a founding partner

The founder on your project: No hand-offs to juniors. The person you talk to is the person who builds.
Pilot pricing: Founding partners get a free or reduced-cost first pilot.
Real priority: With few partners, your project is never at the back of a queue.
A fair exchange: In return, we ask for honest feedback and, if you're happy, a testimonial or case study.

Who you'll work with

Gowrishankar

Founder, Archetech AI Labs

I'm an AI Solution Architect. My work sits at the infrastructure layer: agent systems, AI governance, and the frameworks that help whole teams adopt AI well rather than just use it. I've designed AI adoption programs for engineering teams, built multi-agent development pipelines, and shipped production AI products in healthcare and business automation.

I started Archetech AI Labs to build the foundations companies will need after the AI transition: systems that are secure, auditable and built to last. When you work with us, you work with me directly.

LinkedIn GitHub gowrishankar.in

Questions, answered plainly

You're a new studio. Why should we trust you?

Because you can check our work before you pay for anything. Our open-source projects are public, the first pilot is small and time-boxed, and you own everything we deliver. If the pilot doesn't earn your trust, you walk away with the work.

Do you work with clients outside India?

Yes. We work remotely with teams anywhere, with overlap hours agreed at the start.

Who owns the code?

You do. Code, docs and IP built for you belong to you once delivered.

How do you price projects?

Outcome-based. We agree the deliverable and the price before work starts, so there are no open-ended hourly bills.

Which AI tools and models do you work with?

Claude, OpenAI and open models, and agents including Claude Code, Cursor, Codex and Copilot. We pick what fits the job, not what we're attached to.

How do you handle our data and secrets?

We work with least-privilege access, never copy production data without permission, and can sign an NDA before any detailed discussion.

Is AthenOps available now?

It's in early access. Join the list and we'll reach out when onboarding opens.

Building something with AI? Let's make sure it's built on solid ground.

One 15-minute call. You'll leave with a clear next step, whether or not we work together.

Book a 15-min call