Black Lily: Private On-Premise AI for Hedge Funds, RIAs, and Family Offices
Black Lily installs a complete language model on hardware your firm owns, together with the software your staff uses it through: a chat window, document upload, and search across everything you have loaded. It works the way the AI assistants your team already knows work. The difference is that nothing they type and no file they open leaves your network. Based in Philadelphia, Pennsylvania. Founded by William Dorman, Founder & CEO.
Why it runs on your hardware
- Your files never leave your network. The model runs on a machine in your office, so your documents and everything your staff types into it stay inside the network you already control. No third-party AI provider sits in the path, no outside account holds your files, and nothing your team writes is retained or used to train anyone else's system.
- It works like the tools your team already uses. A chat window, a file upload, and a search box across everything you have loaded. No prompt engineering, no new vocabulary, no training course.
- A data path you can draw on one page. Reviewing a cloud tool means reviewing a vendor: its retention terms, its subprocessors, and a contract that changes when the vendor decides it does. Reviewing this means reviewing one machine already inside your perimeter.
- One deployment, not a per-seat bill. You own the hardware and the deployment, so the tenth person on it costs the same as the first.
- The AI does not depend on anyone else's uptime. The model runs on your hardware, so it keeps answering and everything already indexed stays searchable even with no internet connection. The connectors and the market data feed are network services and catch up when the connection is back.
What is on the box
The model is the part everyone asks about. What your team actually works in is everything around it.
- Your whole corpus, in one place. Everything your firm loads onto it, plus SEC EDGAR filings pulled in and kept current. Ask one question across all of it and get the source passage back.
- Chat that remembers, and is searchable. Every conversation is kept and searchable, so analysis from three months ago is findable rather than lost in a closed window. Searching across colleagues' chats is available as an option, scoped to the groups you choose.
- Projects. A persistent workspace per name, deal, client, or mandate, holding its documents, its chats, and its outputs together.
- Scheduled jobs. A daily run that does the thing someone would otherwise do at seven in the morning: new filings on your names, a digest of what changed, a draft waiting when the first person sits down.
- Dashboards you spin up and hand over. Build a view, share it across the office for a meeting, and it tears itself down on a timer rather than becoming another stale internal page.
- Connectors to the systems you already run. Slack, Box, Dropbox, Microsoft 365, Google Workspace, Salesforce, and SEC EDGAR come standard, so the system reads from where your work already lives.
- Real user accounts and permissions. Named users, groups, and permissions you configure. On-site access is the default; remote access over infrastructure you control is a decision made together at install.
- Live market data, if you want it. Live prices streamed in from a market data vendor over a one-way inbound connection, priced separately. Quotes come in; your documents, your prompts, and your positions do not go out.
Who we build these for
Three kinds of firm with one problem in common: the documents most worth handing to a model are exactly the ones that are not allowed to leave.
- Hedge funds: query an entire diligence data room in one question, summarize broker research and transcripts against your own thesis notes, draft LP letters and DDQ responses from prior versions, and search years of internal research. Confidential material stays inside the perimeter, so an outbound prompt never becomes a data path your MNPI policy has to cover.
- Registered investment advisers: summarize a household's full document history before a review meeting, draft client letters and policy statement updates from the firm's own templates, and search compliance manuals and prior filings. Client information never reaches an outside processor, so your incident response program has no extra surface to cover.
- Family offices: ask one question across trust instruments, entity charts, and partnership agreements, summarize K-1s and manager reports at close, and find the document nobody remembers filing. Records about named private individuals never leave the office.
How a deployment works
A complete open-weight language model installed on a machine in your office, with chat, document upload, and search software on top of it. See the hardware, the models, and the four delivery stages.
- Scope: a 30-minute call, then a walkthrough of where your documents live and which recurring tasks are worth handing to a model. You leave with a part list, a number, and an honest account of what the system would not do.
- Install: we size and configure the hardware, deploy the model and the interface onto it, wire up the connectors you named, index your document stores in place, and set up users and permissions. Nothing is migrated anywhere.
- Train: hands-on sessions with the people who will use it daily, plus written documentation and recorded walkthroughs.
- Maintain: monthly monitoring, model updates as stronger open-weight releases ship, and adjustments as your workflows change.
Typical delivery is four to eight weeks from scoping to daily use, with hardware lead time usually the longest single item.
What we do not claim
Black Lily deploys open-weight models that it did not train. A local deployment does not by itself make a firm compliant with any rule; compliance is your program, and we build to the controls you already run rather than selling a certification. Open-weight models trail the best hosted models on frontier reasoning, and we say which of your workflows fall on which side of that line during scoping rather than after you have bought a machine.
About
Black Lily was founded by William Dorman in 2025 and is based in Philadelphia, Pennsylvania. It exists because the firms with the most to gain from a language model are the ones structurally barred from using one: their documents are privileged, confidential, or personal, and every product on the market asks them to upload it somewhere. No account managers, no layers. You work directly with the founder who scopes the problem, sizes the machine, installs the system, trains your team, and maintains what we deliver.
Contact
Phone: (432) 234-3779
Book a free 30-minute scoping call: cal.com/black-lily/30min
Website: blacklily.ai
Blog