Enterprise Support Center

Support for institutional clients, direct from the people who build it.

Platform Overview

What the terminal integrates with, and the controls around it.

View Overview

Scheduled work

What runs on a schedule, and what is built but not enabled here.

What runs

Security

Our cybersecurity policy, published in full — data handling, access control, encryption, incident response and vendor risk.

Read the policy

Frequently Asked Questions

What's included in a Warren seat?

Full platform access: the complete terminal, every data feed it carries, the assistant, and the APIs. No per-terminal, per-feed or per-API-call fees. Access is quoted per engagement — request access and we will scope it with you.

Can I deploy Warren Data on-premise or in my VPC?

Warren is deployed today as a hosted service running from an immutable container image, so a private deployment inside your own infrastructure is architecturally straightforward. We do not publish a self-service on-premise package, and we would rather say so than list installation artefacts we have not written. If deploying inside your own environment is a requirement, tell us and we will scope it with you.

What latency can I expect end-to-end?

Server-side, a read costs 1–5 ms — a measured in-process p50 of 1.6 ms against a working set of indices, rates, commodities, FX and crypto held continuously warm in memory and refreshed every twelve seconds. End-to-end is the network's figure, not ours: our own capability audit measured 138–190 ms wall-clock from a remote client, and we quote that rather than a number we did not measure. Note what the warm cache is and is not — it makes a figure arrive quickly, it does not make the figure newer, and the market data behind it is delayed and labelled as such on every screen. Questions with a factual answer — a price, a level, a spread, whether a venue is open — are resolved straight from that memory without touching a model, which is why they return in milliseconds rather than seconds. Multi-source research that reads filings before it answers necessarily takes longer; that is a different operation, and the interface tells you which one you asked for. None of these figures is an execution latency: they measure the data path, and the separate Ultra-Low-Latency (ULL) Execution Platform is a different product with its own separately measured numbers. That name is a claim about an execution path and never about the data — the market data does not become faster because a ULL Execution Platform is attached, and no end-to-end execution figure exists, because nothing in that platform has ever sent an order to a venue.

How do I get API access for my trading systems?

Not yet. Warren's data layer is a JSON API internally — quotes, options chains with Greeks, fundamentals from SEC XBRL, filings search, screening — and every response carries its own provenance. What does not exist yet is the product around it: issued keys, per-key quotas and a published specification. It is on the roadmap and we would rather say so than sell you a key we cannot cut. If you want to be first, tell us what you would call and we will build against a real use case.

What compliance standards do you support?

Honestly: audit logging is real, certification is not. Three append-only trails record administrative actions, order intent and agent activity — insert-only, with no delete path, which is what makes them evidence rather than a log. Our full cybersecurity policy is published, including the controls we do not have. We are NOT SOC 2 certified, we hold no SOC 2 report, and we have not signed a BAA or a DPA. If your compliance team needs those, ask us where we are rather than assuming — the answer today is that we would be starting the process with you.

Need Implementation Support?

Our solutions engineers handle onboarding, custom integrations, and performance tuning.

Contact Solutions Engineering