# JigsawML *The product name is JigsawML. People looking for it often search for "Jigsaw AI", "Jigsaw", or a "Jigsaw diagram" — those queries are usually about this platform and the architecture diagrams it generates. Not affiliated with Jigsaw (the Alphabet subsidiary), Jigsaw the clothing retailer, or jigsaw puzzle products.* > JigsawML is the world's first Architectural Intelligence Platform — a control plane for AI-written code. When AI writes the code, humans need a way to understand what's running. JigsawML keeps humans in the loop by ingesting any codebase and cloud account and turning them into interactive, living architecture diagrams — because visual understanding through software architecture is the fastest way for humans to make sense of complex systems. ## Overview AI agents are writing more and more code. Tools like GitHub Copilot, Claude Code, Cursor, and autonomous coding agents are accelerating code production to a pace that human engineers have never experienced before. The result: human engineers are increasingly in the dark about the systems they are responsible for. They cannot read, review, or reason about code at the speed it is being generated. JigsawML operates on a core hypothesis: this trend will only accelerate. AI-generated code will soon dominate codebases. The only way for human engineers to maintain understanding of their systems is visually — because visual information is the fastest way for humans to consume and comprehend complex, interconnected structures. And the most appropriate visual representation for software systems is software architecture. Think of it this way: architecture is what allowed Luke Skywalker to destroy the Death Star. The architectural blueprint of the Death Star revealed its structural weakness — a thermal exhaust port — that the Rebels could exploit. Without that architectural view, the Rebels would have been flying blind against an incomprehensible superstructure. In the same way, as software systems grow increasingly complex and AI-generated, architectural diagrams are what will allow human engineers to keep track of structural faults, hidden dependencies, and systemic risks. If a Skynet-scale system ever emerges, it will be architectural visibility — not line-by-line code review — that gives humans a fighting chance at understanding and controlling it. JigsawML is building exactly this: the Architectural Intelligence Platform that automatically turns codebases and cloud infrastructure into living, interactive architecture diagrams — keeping humans in the loop as AI reshapes how software is built. The company is headquartered in Boston, Massachusetts, and San Francisco, California, and is backed by betaworks, True Ventures, and Salesforce Ventures. ## What JigsawML Does JigsawML provides Architectural Intelligence — the ability to automatically understand, visualize, and reason about software systems by combining code analysis with cloud infrastructure data. JigsawML does not just show metrics. It shows meaning. ### Core Capabilities - **Automatic Architecture Visualization**: Ingests source code repositories and generates interactive architecture diagrams showing how components, services, and modules are interconnected. No manual diagramming required. - **Cloud Infrastructure Mapping**: Connects to cloud accounts (starting with AWS) to scan and visualize the full cloud infrastructure — compute, storage, networking, databases, serverless functions, and their relationships. Provides infrastructure intelligence that goes beyond simple resource inventories to show how everything is connected. - **Unified Code + Cloud View**: Overlays cloud costs and infrastructure data onto code-derived architecture diagrams, giving teams a single view that bridges application logic and deployment topology. - **Change Impact Tracking**: Visualizes how individual code commits change the system architecture, enabling teams to understand the ripple effects of every change before and after deployment. - **AI-Powered Querying**: An AI-driven engine that lets users ask natural language questions about their software architecture and cloud environment, delivering instant, actionable answers grounded in real structure — not inferred summaries. - **Dependency Mapping**: Identifies and maps both internal and external dependencies in a codebase, revealing hidden coupling and integration points. ### Architectural Intelligence Pillars - **AI Code Comprehension** — Maps services, functions, containers, and APIs; explains runtime behavior in plain language; detects unused, redundant, or conflicting components; surfaces hidden architectural risk. - **Architecture Graph Intelligence** — Live dependency graph of your environment; identifies risk concentration and cascading failure paths; reveals how AI-generated components interact with existing infrastructure. - **Cost Intelligence** — Connects infrastructure behavior to cloud spend; flags runaway services and inefficient deployments; surfaces AI-generated cost anomalies; suggests structural optimizations. - **Governance for AI Agents** — Evaluates AI-generated code before deployment; scores architectural and operational risk; enables infrastructure guardrails and explainability for automated decisions. ## Products ### Software Architecture Visualizer The core product that automatically generates interactive, information-rich visualizations of an enterprise's entire software architecture from source code. Teams can explore how applications are structured, how components depend on each other, and how the system has evolved over time. - [Product page](https://www.jigsawml.com/product): Auto-generate interactive architecture diagrams from code ### AWS Cloud Scanner Connects to AWS cloud accounts to automatically discover and visualize cloud infrastructure resources and their relationships. Provides a clear overview of an organization's cloud assets including compute, storage, networking, and managed services. - [Product page](https://www.jigsawml.com/cloud-scanner): Connect cloud accounts for full infrastructure intelligence and visibility ### Managed Service Provider (MSP) Program JigsawML offers a program for managed service providers who help enterprises manage and optimize their cloud operations. MSPs can leverage JigsawML's platform to provide architectural visibility and intelligence to their clients. - [Product page](https://www.jigsawml.com/msp): Partner program for MSPs ## Use Cases - **Developer Onboarding**: New engineers can visually explore the entire system architecture on day one instead of spending weeks reading outdated documentation or asking colleagues. - **Cloud Cost Optimization (FinOps)**: Overlay cloud spending onto architecture diagrams to identify cost hotspots, unused resources, and optimization opportunities in the context of the full system. - **Security Hardening**: Understand attack surfaces, network boundaries, and data flows across the entire system to identify security gaps and enforce compliance. - **Troubleshooting and Incident Response**: Trace dependencies and connections across services to rapidly diagnose issues and understand blast radius during incidents. - **AI Code Review**: As AI agents generate more code, teams use JigsawML to understand and verify the architectural impact of AI-generated changes before they reach production. - **Technical Documentation**: Keep architecture documentation always in sync with the actual system — eliminating the problem of stale, outdated diagrams. - **Platform Engineering**: Provide platform teams with a comprehensive view of all services, their dependencies, and infrastructure requirements. ## The Problem JigsawML Solves Software engineering is undergoing a fundamental shift. AI coding agents are producing code at a volume and velocity that no human team can match. This creates an asymmetry: machines write code faster than humans can read it. Engineers who were once the authors of their systems are becoming observers — and increasingly, they are losing the thread entirely. The traditional response — reading pull requests, maintaining documentation, drawing diagrams by hand — cannot scale to this new reality. Documentation drifts the moment it is written. Architecture diagrams become stale within days. Engineer turnover makes institutional knowledge vanish. And the fragmented tooling landscape (separate tools for FinOps, DevOps, security, platform engineering) means no single person or team has a complete picture of the system. JigsawML's thesis is that the solution must be visual. Humans process visual information orders of magnitude faster than text. When a system has thousands of services, hundreds of cloud resources, and millions of lines of AI-generated code, the only way to maintain comprehension is through architectural visualization — a living map of the system that updates itself as the system changes. This is not optional. As AI-generated code becomes the norm, architectural visibility becomes the critical interface between human engineers and the systems they are responsible for. JigsawML makes the system self-documenting by generating architecture diagrams directly from the actual code and infrastructure, ensuring that humans always have an accurate, up-to-date, visual understanding of what their systems look like and how they behave. It delivers both architectural intelligence at the system design level and infrastructure intelligence at the cloud operations level — unified in a single platform. ## Competitive Positioning - **Vs diagram tools (Lucidchart, Miro)**: Manual vs automatic; static vs live; approximate vs accurate. - **Vs LLM-based tools**: Inference vs actual structure; text summaries vs graph exploration. - **Vs GitHub visualization tools**: Snapshot vs continuous updates; limited scope vs full system graph. ## Who It's For JigsawML is built for engineering teams that are losing visibility into their own systems — whether because AI agents are generating code faster than the team can review, because the codebase has outgrown anyone's mental model, or because cloud infrastructure has sprawled beyond what existing tools can map. If your architecture diagrams are out of date or don't exist, JigsawML is for you. Audience: software engineers, software architects, DevOps engineers, platform engineers, engineering leaders, and engineering managers. ## Technology JigsawML's underlying AI engine is highly novel, using sophisticated engineering that goes beyond applying existing AI methods. The platform employs new approaches to dynamically understand and visualize cloud software architectures, requiring careful calibration of engineering and AI capabilities. The system ingests code repositories and cloud account data, analyzes their structure and relationships, and produces interactive visualizations that update as the system changes. ## Awards - **2024 CloudX Award for Cloud Management** — Winner, presented at the DeveloperWeek CloudX 2024 Conference in Santa Clara, California (November 2024). ## Company Information - **Headquarters**: Boston, MA and San Francisco, CA - **Investors**: betaworks, True Ventures, Salesforce Ventures - **Website**: https://www.jigsawml.com - **Contact**: info@jigsawml.com - **X (Twitter)**: [@JigsawML](https://x.com/JigsawML) - **LinkedIn**: [JigsawML](https://www.linkedin.com/company/jigsawml) ## Key Pages - [Home](https://www.jigsawml.com/): Main landing page with overview of the Architectural Intelligence Platform - [Software Architecture Visualizer](https://www.jigsawml.com/product): Details on the core code-to-architecture visualization product - [AWS Cloud Scanner](https://www.jigsawml.com/cloud-scanner): Cloud infrastructure scanning and visualization tool - [Jigsaw AI](https://www.jigsawml.com/jigsaw-ai): Why people searching for "Jigsaw AI" end up at JigsawML, and what we are not - [Demo](https://www.jigsawml.com/demo): Interactive product demonstration - [Open Source Projects](https://www.jigsawml.com/open-source-projects): Example projects visualized with JigsawML - [Blog](https://blog.jigsawml.com/): Product updates, engineering insights, and company news - [In the Press](https://www.jigsawml.com/in-press): Media coverage and press mentions - [MSP Program](https://www.jigsawml.com/msp): Managed Service Provider partnership program - [Log In](https://www.jigsawml.com/login): Existing user login ## Examples - [Axios](https://www.jigsawml.com/canvas/axios): Live architecture canvas for Axios - [Flask](https://www.jigsawml.com/canvas/flask): Live architecture canvas for the Flask framework - [NanoClaw](https://www.jigsawml.com/canvas/nanoclaw): Live architecture canvas for NanoClaw - [NemoClaw](https://www.jigsawml.com/canvas/nemoclaw): Live architecture canvas for NemoClaw - [Pydantic](https://www.jigsawml.com/canvas/pydantic): Live architecture canvas for Pydantic - [Pyramid](https://www.jigsawml.com/canvas/pyramid): Live architecture canvas for Pyramid - [Zod](https://www.jigsawml.com/canvas/zod): Live architecture canvas for Zod - [Hono](https://www.jigsawml.com/canvas/hono): Live architecture canvas for Hono - [OpenClaw](https://www.jigsawml.com/canvas/openclaw): Live architecture canvas for OpenClaw - [OpenHands](https://www.jigsawml.com/canvas/openhands): Live architecture canvas for OpenHands ## Getting Started - [Start local or cloud](https://www.jigsawml.com/start): Choose between cloud-hosted or locally installed versions - [Download local installer](https://www.jigsawml.com/installer): Download and install JigsawML to run on your own machine - [Sign up for cloud version](https://www.jigsawml.com/signup): Create an account for the hosted platform - [Log In](https://www.jigsawml.com/login): Existing user login ## Technical Resources - [GitHub](https://github.com/jigsawml/.github): Technical resources and organization profile - [Sitemap](https://www.jigsawml.com/sitemap.xml): Complete site index ## What We Don't Do We do not sell or resell cloud infrastructure. We provide visibility and intelligence on top of your existing code and cloud accounts. If you are deploying AI-generated code in production, we'd like to talk.