VeraGen Hub

Research you can put your name on.

The complete workspace for evidence-based research. AI that reads only the sources you chose, automation for the work that repeats, a full account of what every run did and cost, and a public profile at your own handle. One product, built to a standard the work deserves.

Invitation only. We work with a deliberately small group of researchers.

Grounded, not guessedAnswers come from the sources in your project. Nothing else is in scope.
Traced end to endEvery run records each step, what it read, and how long it took.
Priced to the tokenCost measured from real provider usage, never estimated afterwards.
Published as yoursFindings go out under your own handle, on a profile you control.

The difference

General-purpose AI answers from everywhere. Your client is paying for somewhere.

Behind every question a client asks is a second one they may never say out loud: how do you know that?

A general assistant cannot answer it. It draws on everything it has ever read, blends it, and hands you fluent prose with no way back to a source. Impressive for a first draft. Worthless the moment someone senior asks where a number came from.

The Hub inverts the model. You decide what the AI may read — the papers, pages, datasets and notes you gathered for this piece of work — and every answer stays inside that boundary. Then the platform records what it did. When you are asked how you know, you show them.

That inversion is not a setting or a mode. It is the architecture, and it runs through all six parts of the product.

Your reputation is the product. The work has to survive being checked.
  • Grounded by default. Answers come from the sources in your project, not the open web.
  • Traceable. Every automated run keeps a step-by-step record of what it read and produced.
  • Accountable. Real token counts, real cost, attributed to the workflow that spent it.
  • Yours. Your sources, your outputs, your profile, under your own handle.

A general AI assistant

  • Reads the whole internet, including whatever it got wrong
  • Gives you an answer with no route back to a source
  • Forgets the engagement the moment you close the tab
  • Cannot tell you what it did, only what it concluded
  • Bills you a monthly number with no breakdown behind it
  • Leaves the finished work in a chat log

VeraGen Hub

  • Reads the sources you approved for this project, and nothing else
  • Records which documents were consulted for the answer you got
  • Keeps every source, output and prototype together, permanently
  • Traces every step of every run — timing, model, tokens, outcome
  • Attributes real measured cost to the workflow that spent it
  • Publishes the finished work to a research profile at your handle

How the work moves

Gather, work, publish.

The whole arc of a piece of research in one place — replacing the folder of downloads, the chat window and the document you paste between with a single system that remembers all three.

01 — GATHER

Bring the evidence in

Crawl an entire site into a project in seconds. Upload documents, spreadsheets and datasets. Every source for a piece of work lands in one place, with its own context and reusable skills attached.

02 — WORK

Think with it, not around it

Draft in Studio with an AI partner that already knows your sources. Interrogate the material and get sourced answers. Hand the work that repeats to agentic workflows that run on demand or on a schedule, without you.

03 — PUBLISH

Put your name on it

Publish findings and interactive prototypes to a research profile at your own handle — a serious, permanent home for your work that anyone can read without an account.

What's inside

Six parts. One workspace. No seams.

Not a bundle of tools that happen to share a login. Every part is built for the same job — getting defensible research out of your head and in front of the people who asked for it — and each one knows what the others did.

Built to be checked

Every run leaves a record.

Automated research is only worth having if you can audit it afterwards. Every workflow run in the Hub keeps a full trace: which step ran, what it read, how long it took, which model answered, and what it cost — down to the token.

That record turns "the automation produced this" into something you can stand behind, and turns a surprise on your bill into a question with an answer. Most AI tools cannot tell you what they did. The Hub cannot forget.

  • Per-step timing so you can see where a run actually spends its time.
  • Measured cost, taken from real usage rather than estimated after the fact.
  • Failures you can diagnose — the step that failed, and what it was given.
Observability — last 30 days
Runs
128
workflow executions
Succeeded
124
96.9% clean
Avg step
1.7s
across all nodes
AI spend
$18.40
measured, not estimated
chainworkflow.run2 steps · 2.7s
trigger.manual41ms
source.project590ms
agent · sonnet1.7s
output.project210ms
llm1,842 in · 356 outclaude-sonnet · $0.011

Who it's for

Built for people whose name is on the work.

Not for casual questions. For the research someone else will act on, question, and hold you to.

Independent researchers

You produce findings other people act on. The Hub keeps the evidence and the output together, so the work is still defensible a year later when someone reopens it.

Consultants and analysts

Client work lives in its own project with its own sources. Nothing bleeds between engagements, and nothing is answered from outside them.

Small research practices

Shared account credit, per-member limits and roles, so a small team can work together without anyone quietly running up the bill.

People building a public body of work

A research profile at your own handle turns finished work into visible credibility rather than a PDF in someone's inbox.

The Hub is ready. Access is by invitation.

We work with a deliberately small group of researchers — close enough that the Hub is built around how the work is actually done. Tell us what you research and we'll be in touch.

Request an invitation