Your environment
Connect from Claude, ChatGPT, or any client that supports the Model Context Protocol. There is no new interface to learn and nothing to install alongside your existing tools.
Early access
Work with HuxleyIQ from inside your own AI assistant.
The HuxleyIQ connector speaks the Model Context Protocol — an open standard that lets an AI assistant reach an outside system on your behalf. Connect it once and your HuxleyIQ workspace becomes something you can ask for by name from Claude, ChatGPT, or any MCP-compatible client: the deal, the completed analysis, the tasks it was broken into, the research behind it, the source document. You stay in the tool you were already working in.
What it is
Your workspace, reachable from wherever you actually write.
Diligence output has to end up somewhere else — an IC memo, a model, a call, a note to a capital partner. The connector removes the copy-and-paste step between HuxleyIQ and that destination.
One source of truth
HuxleyIQ remains the system of record for the deal, the question set, and the firm memory behind both. The connector is a window onto that workspace, not a second copy of it.
Read-only, on purpose
The connector pulls. It never pushes. Nothing your assistant does through it can alter a deal, a document, or an analysis in your workspace.
What you can pull
Finished work, not raw search results.
HuxleyIQ analyses are long. A single investment analysis produces a primary report, the research artifacts underneath it, and a task list showing how the work was broken down. The connector is built to hand you the specific piece you asked for rather than everything at once.
Your deal list
Ask for active deals, the high-priority ones, or a deal by name. Each comes back with its status, priority, deal type, who is leading it, and when it last moved.
A single deal in full
Everything on file for one opportunity: the documents uploaded to it and every analysis run against it, each with its own status, so you can tell finished work from work in progress.
Completed analyses
The primary report a deal team reads, plus the research artifacts behind it — competitive landscape, financial analysis, valuation, risk analysis, market research, value creation, and the diligence question set.
A section, not just a file
An investment report can run past ninety thousand characters. Ask for the valuation section or the risk section by name and that is what you get, at a length your assistant can actually reason over.
The tasks behind a run
Every task a workflow was broken into. Your assistant can follow how a conclusion was reached rather than only reading the conclusion — and see what the analysis set out to check.
Source documents
The CIM, the financials, the teaser, and the data-room files attached to a deal — listed with their extraction status, and readable in full, so your assistant quotes the source instead of paraphrasing a summary of it.
How it works
Four steps, once.
-
Connect
Add the HuxleyIQ connector in your client and sign in to your workspace. Setup instructions live behind your HuxleyIQ login.
-
Ask
Ask in the words you would use with a colleague — "what did our analysis say about customer concentration on the Midwest distributor?"
-
Pull
Your assistant locates the deal, finds the relevant analysis, and retrieves the section that answers the question.
-
Work
Draft the memo, build the slide, pressure-test the assumption — in the document you were already in, against what HuxleyIQ actually found.
Better together
The deal truth, sitting next to everything else you have connected.
Most teams adopting MCP do not connect one system. They connect the inbox, the meeting recorder, the CRM, the task tracker — and then find the assistant can only reason about the surface of a deal, because the analysis itself lives somewhere it cannot reach. This is the piece that closes that gap.
Connected through MCP
HuxleyIQ
Deals, completed analyses, the tasks and research artifacts behind them, and the source documents they were built from — held to your firm's question set.
The thread with the banker or the founder: what was sent, what was promised, and what is still outstanding.
Meeting recordings
What management actually said on the call — in their words, with the hedges and the qualifiers intact.
CRM & task management
Who owns which workstream, where the process stands, and what has to close before the committee meets.
Your AI assistant
One place to ask a question that crosses all of them. MCP is what lets the assistant reach each system on your behalf, so the answer is assembled from every source at once instead of from whichever one you happened to paste in.
What that makes possible
Claims you can cross-check
"Management said retention improved on the call — what did our analysis find, and does the CIM say the same thing?" Three sources, one question.
Drafts that arrive with the context
An IC memo, a capital-partner update, or call prep drafted against the real analysis and the real correspondence, not a summary of either.
Follow-ups that land somewhere
Open diligence questions become the tasks your team already tracks, assigned in the system they already use.
A note on direction. The HuxleyIQ side of this picture is read-only — your assistant pulls from your workspace and cannot write back to it. Whether it can write to your inbox, CRM, or task tracker depends on the connectors you have granted those systems, not on this one.
Portability, not lock-in
We would rather a firm stayed because the product compounds than because leaving is painful. A read-only connector is the plainest form of that: the analysis your team paid for is reachable from anywhere you work, in a standard open protocol, without asking us for an export.
How we handle your dataFAQ
The questions deal teams ask first.
Can we use HuxleyIQ from Claude or ChatGPT?
Yes. The HuxleyIQ MCP connector lets a client that supports the Model Context Protocol — such as Claude or ChatGPT — reach your HuxleyIQ workspace and pull deals, completed analyses, research artifacts, workflow tasks, and source documents into whatever you are working on. The connector is read-only and is available in early access on request.
What is the Model Context Protocol?
The Model Context Protocol (MCP) is an open standard for connecting AI assistants to outside systems. A system publishes a small set of capabilities, and any MCP-compatible assistant can use them on a user's behalf without a custom integration being built for each pairing.
Can the HuxleyIQ MCP connector change anything in our workspace?
No. The HuxleyIQ connector is read-only. It can list and read deals, analyses, artifacts, tasks, and documents; it cannot create, edit, or delete anything, and it cannot start or stop an analysis.
Who can see what through the HuxleyIQ MCP connector?
Access mirrors the permissions the connecting user already holds in HuxleyIQ. The connector reads a workspace as that user, so it cannot return a deal or document that person could not open in HuxleyIQ directly.
Does using the connector mean our deal materials go to our own AI provider?
Only what a user pulls, when they pull it. Content retrieved through the connector enters the AI environment your firm has chosen and is then governed by your firm's agreement with that provider. HuxleyIQ keeps customer deal materials confidential and does not use them to train models.
Is the HuxleyIQ MCP connector available now?
The HuxleyIQ MCP connector is in early access. Firms can request it, and setup takes a short conversation with the team rather than a self-serve install.
Early access
Bring your own environment.
Tell us which client your team works in and what you would want to pull first. We will get the connector switched on for your workspace and walk through it on a live deal.
Request the connector
Pick a time to meet with a HuxleyIQ founder.
Meet with a founder founders@huxleyiq.com