Gerolamo Technology Corpus|98,262 Sources

Proof

We build real capability and ship it to production.

Everything on this page we designed, developed, and distributed ourselves. Some open source, some proprietary, all running in production.

Here is everything we have built.

The full lifecycle, proven.

Architecture

System design, data models, integration surfaces, and technical decisions that hold up under real production load.

Narrative design

Positioning, vertical go-to-market, and the language that makes complex technology land with buyers.

Pricing models

Freemium tiers, enterprise licensing, engagement-based structures, and open source strategies that create adoption and conversion.

Distribution

MCP servers, APIs, npm packages, web applications, and federal marketplace positioning. Every channel designed for the capability it serves.

What we've built

Everything we've designed, developed, and distributed.

Everything here we built for ourselves. These are the tools that automate our own delivery and the proving grounds for our cryptographic patterns. Each one went through the same loop we run for clients, and product pages and repositories are linked where they are public.

This is a sample, not the whole of our work. Our client delivery has supported consequential missions and institutions under agreements that keep them off this page.

Technical Intelligence

Gerolamo

Production SaaS

A technical intelligence engine that scores over 50,000 entities across GitHub, arXiv, and Hugging Face on defensibility, velocity, and frontier risk. We run it to de-risk how we design, develop, and distribute R&D: it tells us what is real, what is moving, and what a frontier lab is about to absorb, before we spend build time on it. Delivered as an MCP server, API, and web app.

Our design work

Scoring methodology, ingestion pipeline, freemium-to-pro pricing, and MCP, API, and web distribution.

Agent Harness

Ultimus

Proprietary / Local-first

Our agent harness. A local-first agentic coding system that runs the same loop against any model, hosted or fully local, with permission-gated tools, sandboxed execution, syntax-tree-validated edits, and tiered memory. It benchmarks near Claude Code on open-source models. This is what generates software and pushes it to production.

Our design work

Harness architecture, model router and gateway, permission and sandboxing model, and tiered memory.

Artifact Harness

The Production Harness

Proprietary / Internal

Our artifact harness. Ultimus builds the software; this builds the go-to-market. It encodes brand, voice, design tokens, and governance so decks, proposals, capability statements, and outreach ship fast, on-brand, and pre-validated. Every artifact moves a business outcome, which is how one team ships four to six fundable artifacts a week.

Our design work

Skill architecture, brand encoding, quality gates, and an artifact model where each output feeds the next.

Agent-native CRM & Ops

Workplay

Proprietary / Internal

An agent-native CRM that is aware of our R&D. It works as a CRM and a task-management platform in one, software-defined so agents read from and write to it directly, and it integrates the rest of our stack. Native SAM.gov and USAspending MCP feeds pull federal signals straight into the pipeline.

Our design work

Hub-and-spoke data model, agent-readable task and pipeline layer, MCP integration strategy, and federal signal feeds.

Governed Data Collection

Sonde

Open Source

An MCP-native topic registry for monitoring GitHub, arXiv, Hugging Face, RSS, and web sources. Its declare, simulate, collect pattern gives agents governed access to fresh data with a full audit trail.

Our design work

Collection governance framework, simulate-before-production pattern, lineage model, and developer experience.

Visual Evidence Capture

Snap

Open Source

Structured screen capture that feeds visual context straight to AI agents over MCP. Annotate any screen and hand an agent the evidence, with no custom integration per tool.

Our design work

Annotation schema, MCP tool interface, and workflow integration for operations review and analysis.

Security Intelligence

Honeymoon

Open Source

Multi-agent security scanning with cryptographic proof of every action. The scan, simulate, harden, sign loop runs locally and airgapped, so you get a verifiable security posture without sending data to a third party.

Our design work

Security workflow architecture, cryptographic attestation model, and airgap-first distribution.

Cryptographic Signing & Verification

Zephyr

Proprietary

Environment-agnostic signing and policy enforcement for high-consequence data. Envelope-based signing, a Gatekeeper policy agent, and tamper-evident records of every high-consequence action, with vertical narratives across healthcare, energy, defense, logistics, and science.

Our design work

Envelope signing model, Gatekeeper policy agent, five vertical go-to-market narratives, and DoD Tradewinds positioning.

Codebase & Document Intelligence

Prelude

Open Source / npm

Reads a codebase or a document set and produces a structured, machine-readable account of what is in it: stack, architecture, dependencies, decisions, and changes. The input layer that grounds agents in what is actually there. Open source on npm with an MCP server.

Our design work

Context schema, npm distribution, and MCP server interface.

Local-First Document Signing

Inksign

Proprietary / Desktop

A local-first PDF signing tool that produces tamper-evident PAdES signatures while private keys never leave the machine. It signs hand-drawn documents, converts Word and Markdown to PDF, and merges or splits files, with browser-based verification that needs no install. A proving ground for our signing and verification patterns.

Our design work

Local-key signing architecture, PAdES implementation, cross-distribution packaging, and no-install browser verification.

Verifiable Publishing

Progoff

Proprietary / Web

A digital publishing platform built as a private press for the inner life, with essays on privacy, technology, and personal data. Every piece is cryptographically verifiable through SHA-256 content hashes and served in agent-readable formats including MDX, JSON-LD, and feeds. A testbed for verifiable content and dual human and agent distribution.

Our design work

Editorial model, SHA-256 content verification, and dual human and agent-readable distribution formats.

Why this matters for your R&D.

Every capability here started as raw R&D: an idea, a prototype, a technical insight that was not yet shaped for a buyer. We took each one through the same loop: isolate the features that matter, break larger systems into consumable offers, and position against real demand.

The same process we run on our own work applies to yours.

When we say we can turn your research into revenue, this page is the evidence.

Your R&D deserves the same treatment.

We'll design the capability, develop it, and distribute it to the people who need it. Sixty minutes, free. Let's talk about what you've built.