ShadowCaster
Corporate Digital Twins
Emails reflect your processes.
Let us show them to you.
ShadowCaster builds a digital twin, a living map, of your organisation using the information embedded in email content.
People, departments, buildings, partners - it's all in there.
A knowledge graph that reflects the true shape of your business providing intuitive observability and the firm foundation for AI automation. And not a byte of it has to leave your infrastructure.
- Local extraction
- Local inference
- Observe
- Automate
- Own
From the team behind data platforms at F1, Olympic broadcast & investment banking · Trading since 2009
Every organisation runs on email. Projects are commissioned, decisions are made, commitments are recorded, and relationships are built — all in the inbox. But none of that knowledge is structured, none of it is connected, and almost all of it is forgotten the moment the thread scrolls off screen. Your CRM captures what sales chose to type in. Your project tool captures what someone remembered to update. Email captures what actually happened. ShadowCaster turns that into something you can use.
Extract. Connect. Act.
Three layers, one pipeline. From raw email to a live knowledge graph to automated action — without your data ever leaving your control.
ShapeExtractor
A local agent that reads your email and runs a language model on your infrastructure to extract entities, relationships and commitments. People, projects, organisations, deadlines — pulled from unstructured text and written to a local store. Nothing leaves your network.
- Runs on-prem or on your laptop
- Local LLM inference
- Privacy by architecture
ShadowStore
The structured knowledge graph that ShapeExtractor builds. Every entity is connected to every other entity it touches: who works on what, who committed to whom, which projects depend on which people. A living model of your organisation that updates with every email.
- Entity-relationship graph
- Temporal: tracks state over time
- Queryable via API
ShadowTasks
An event-driven automation layer that watches the graph and acts on it. When a commitment is approaching its deadline, when a project has gone quiet, when a key person has dropped out of a thread — ShadowTasks surfaces it before it becomes a problem.
- Proactive observability
- Event-driven automation
- The digital twin that works for you
The extraction problem just became solvable.
Organisations have always known that their email contains institutional knowledge. The reason nobody has built this before is that extracting structured data from unstructured text, reliably, at volume, was not economically viable until small language models became good enough to run locally. That happened in the last eighteen months.
ShadowCaster exists at the intersection of two things: the engineering to build data platforms that handle volume and velocity (which we have spent twenty-five years doing) and the AI capability to extract meaning from text without sending it to a third party (which is now, finally, possible). The window is open and will not stay open long.
Built by a team that has done this before, at scale
The founder built and operated a 10+ petabyte media archive for a global motorsport organisation, live broadcast monitoring systems at the Olympics, and document platforms in investment banking. The hard problem in ShadowCaster is the same hard problem in all of those: a data path from messy source to clean, queryable, reliable store, running in production under real conditions. Different domain. Same discipline.
We have built this kind of system before.
ShadowCaster is a new product, but the engineering underneath it is not. The same team has built and operated data platforms across regulated finance, live broadcast, elite sport and global education — always at scale, always in production.
- 25+
- 10+ PB
- 20+
- 2009
Years designing and operating production data platforms
Media archive built for a global motorsport organisation
Live broadcast feeds monitored at the Tokyo Olympics
Trading since. Finance, broadcast, sport, education.
Interested in what ShadowCaster can see in your organisation?
We are building ShadowCaster now. If you run a company where institutional knowledge lives in email and you want to be an early design partner, or if you invest in AI-native enterprise tools and want to understand the thesis, we would like to hear from you.
- rdd@palefire.io
- Studio
- London · Remote worldwide