Our

Our Products

Real software I've built and shipped — not concepts

Built By Me, Not A Team

Production Tools I've Built and Shipped

Everything below is running in production today — three of them open source on GitHub, one licensed to agencies and freelancers. No mockups, no roadmap slides — just what I've actually built, with the real engineering behind it.

  • 4Production Tools Built
  • 99MCP Tools Across mcp-figma & mcp-jira-api
  • 13Servers in the MCP Ecosystem
  • 233Automated Tests on the Marketing Platform
Open Source · MIT

Tech Stack

Python (stdlib) MCP / JSON-RPC Figma REST API Figma Plugin API TypeScript
Product 01

mcp-figma

An MCP server that turns a real Figma file into working code — automatically

Normally, closing the gap between a Figma file and real code means a human eyeballing colors and spacing and hand-translating them by feel. I built mcp-figma to remove that step — it gives Claude Code direct, structured access to a Figma file's actual content over the Model Context Protocol, using nothing but Python's standard library, no external HTTP dependencies.

It reads real design tokens, checks accessibility contrast automatically, generates starter component code in five different frameworks, and ships with a companion Figma plugin that can build a file back from a design spec — so design and engineering stay in sync without anyone doing the translation by hand.

Why It's Worth Noting

  • 47 tools across 8 domain modules — file inspection, design tokens, webhooks, accessibility, and code generation
  • Extracts real design tokens and exports them to the W3C DTCG spec, resolving alias chains with topological sort
  • Generates starter component code for React, SwiftUI, Android, CSS, and Flutter straight from a Figma frame
  • Computes both APCA and WCAG 2.1 contrast, and can search a palette for accessible color pairings
  • Companion Figma plugin builds files back from a design spec — 121 tests, 100% code coverage
View on GitHub
Open Source · MIT

Tech Stack

Python (stdlib) MCP / JSON-RPC Jira Cloud REST v3 Jira Server REST v2
Product 02

mcp-jira-api

A Jira server that runs the full issue lifecycle — and reasons like a Scrum master

Most Jira tooling only shows you raw sprint data. I wanted something that could actually manage the full issue lifecycle for Claude Code — create, transition, link a PR, close — and go a level deeper: turn that same data into the kind of statistical insight a good Scrum master would give you, without a human doing the math by hand.

mcp-jira-api runs against both Jira Cloud and Jira Server/Data Center from the same codebase, and its analytics layer is built on real statistics — confidence intervals, Markov chains, Monte Carlo forecasting — not just averages and bar charts.

Why It's Worth Noting

  • 52 MCP tools spanning issue lifecycle, ceremonies, agile analytics, and release management
  • Real statistics, not guesswork: Bootstrap confidence intervals for velocity, PERT estimation, Monte Carlo delivery forecasts
  • Team health scoring built on Spotify's Squad Health Check and Edmondson's Psychological Safety Scale
  • India-specific layer: IST capacity correction, holiday-aware forecasting, NASSCOM AgileX maturity mapping
  • JQL-injection safe and STRIDE-reviewed with zero residual security findings
View on GitHub
Open Source · MIT

Tech Stack

Python (stdlib) mxGraph XML AST / Call-Graph Analysis
Product 03

mcp-drawio-diagram

Architecture diagrams generated straight from your code — and fully editable

Most auto-generated architecture diagrams are one of two things: a static image you can't touch, or a hand-drawn one a developer forgets to update. I built mcp-drawio-diagram to statically analyze a codebase's real AST and call graph and turn that into actual draw.io files — diagrams that open fully editable, generated from what the code really is.

It covers all 12 standard UML diagram types with no external rendering service, no Graphviz, nothing beyond pure mxGraph XML generated locally — so a team's system documentation can be regenerated from the real codebase whenever it's needed, instead of slowly rotting out of date.

Why It's Worth Noting

  • All 12 standard UML diagram types, generated locally with zero external rendering dependencies
  • OMG UML 2.5–verified connectors — composition, aggregation, realization, and dependency drawn correctly, not reused from generic ERD arrows
  • Files open fully editable in draw.io desktop, app.diagrams.net, and the VS Code draw.io extension
  • Auto-generates shareable app.diagrams.net collaboration links
  • Can regenerate editable draw.io files directly from existing Mermaid diagrams, no re-analysis required
View on GitHub
Paid Product
Claude Marketing Tool dashboard screenshot

Integrations

Google Ads API Meta Ads API LinkedIn Ads API
Product 04

Claude Marketing Tool

An agency-grade audit-to-ad-campaign pipeline, running end to end in minutes

A proper SEO and content audit, competitor research, a campaign strategy, and platform-specific ad copy — that's the kind of engagement a marketing consultant or agency normally charges thousands of dollars for, and it usually eats hours of an analyst's time per client. I built this to automate that entire pipeline for close to zero marginal cost.

Give it a URL, and within about 60-90 seconds it comes back with a scored, professional audit report and a full campaign strategy. From there it can write ad copy for Google, Meta, and LinkedIn — and, if you want, push and manage those campaigns live from the same dashboard.

Why It's Worth Noting

  • Full 8-phase audit — crawl, keyword research, 12-month trend analysis, competitor discovery, scoring — in roughly 60-90 seconds
  • Generates real ad copy for Google Ads, Meta Ads (5 copywriting frameworks), and LinkedIn Ads
  • Can push and manage live campaigns on Google, Meta, and LinkedIn, including audience building and retargeting
  • Full web dashboard: bulk audits, campaign management, PDF export, white-label branding, role-based accounts
  • Production-grade, not a prototype: 233 automated regression tests, SSRF-guarded crawler, CSRF protection
  • Built to resell — this level of automation typically supports $2,000–$5,000 per audit-and-strategy engagement, or $500–$5,000/month on retainer
All Reports — real audits for triglav.in and medspyhealthcare.in, my own client sites.
All Reports — real audits for triglav.in and medspyhealthcare.in, my own client sites.
Report Overview — technical, on-page, content, and AI-readiness scores in one view.
Report Overview — technical, on-page, content, and AI-readiness scores in one view.
Request Access

No self-serve checkout yet — message me and I'll set up access personally.

Let's Build Something Real

Want something built
with this level of care?

These four products are proof of how I work — precise, tested, and shipped, not just planned. If you need something built the same way, let's talk.

Start a ConversationSee My Services