A simulation you can read, and a decision you can record.
This page does three things. It walks through the Revercue thread in eight steps. It offers a
clearly labelled, self-contained simulation of the six main screens, filled with synthetic
data and running entirely in your browser. And it gives you a workspace to write the answer a
user would need, record a recommendation with its evidence, and produce a feedback report.
This is not the production application. No account is created, no Revercue
server is contacted, no email is sent and no payment is taken. The client, the documents, the
amounts and the activity are invented for this page. The simulation reproduces the shape of
the workflow — the states, the log entries, the arithmetic — and nothing about the
production API, database or client pages can be inferred from it.
SimulationSynthetic workspace4 active documents · 3 clients · 1 follow-up due · $0 revenue this monthReset restores the starting state.
1. The simulation
Six screens, one thread. The data starts in a fixed state so two runs of the demo are
comparable: one client (Mara Ellis at Northlight Studio), a published proposal, a contract
waiting on a decision, a form already answered, and an unpaid invoice. Every action you take
appends a line to the communication log and, where it should, changes a figure on the
dashboard. That connection is the product's claim, and the only thing this demo is trying to
show.
JavaScript is off, so the simulation is shown as a description. The tutorial,
the evaluation rules and every explanation below are complete in this HTML. The interactive
panels, the copy buttons and the feedback report need JavaScript; enable it to use them. In
the starting state the workspace shows: 4 active documents, 3 clients, 1 follow-up due, $0
revenue this month; a proposal at /p/northlight-refresh with three packages
(Foundation $4,800, Signature $7,200, Full thread $9,600); a contract awaiting a decision; a
completed requirement form; and an unpaid invoice INV-2026-014 for $7,200 plus $576 tax.
Simulationrevercue.example / demoSession 1 · 10:20 · fresh state
Workspace
Counts are synthetic and do not represent a real account.
Dashboard
SyntheticMara Ellis is the only client shown in detail
4Active documents
3Clients
1Follow-ups due
$0Revenue this month
$7,776 pending across 1 unpaid invoice.
Suggested next action
Pattern, not a modelProposal opened twice, no reply after three days.
Send a short follow-up about the scope, or offer the Foundation package.
“Do it” only opens the right screen. It never creates, sends or charges anything.
Recent activity
Proposal — Northlight Studio site refresh
Public link /p/northlight-refreshStatus: awaiting decision
The client side of the same document. No account, no app — a link with three
packages and a question box, exactly as a real client page presents them.
No package selected yet.
What this screen proves
That a package choice and a question can be structured data rather than an email,
and that both are recorded against one client. It does not prove that the
production proposal renders identically, that tracking is accurate, or that email
delivery works.
Contract — Signature package
Public link /c/northlight-signatureStatus: awaiting decision
The contract inherits the client, the scope and the accepted price. The client has
three actions, and the product records which one was taken.
Parties and scope
Northlight StudioSignature package$7,200
Scope, payment schedule, timeline, revision rounds, cancellation and liability
sections are on the page above the buttons. In the production flow the freelancer
signs this version first, then the client types a full name and ticks consent; both
signatures are recorded with the date. The demo stops at the decision and does not
reproduce the signature step.
No action taken yet.
After approval, this contract is set to continue into the requirement form and then
the invoice.
What this screen proves
That a decision stops being a reply in an inbox and becomes a state on the thread,
with a next document attached. It does not reproduce the signature step or prove
legal sufficiency.
Requirement form
Public link /f/northlight-briefStatus: not submitted in this session
Two questions instead of seven, to keep the demo short. In production the form
supports short text, long text, file upload, single choice, multiple choice, combined
and dropdown, with required flags and optional branching on Growth and above.
What this screen proves
That the brief can arrive as structured answers attached to the same client, ready
to be assembled into a paragraph. It does not prove upload handling, branching
logic or the Prompt Creator, none of which run here.
Invoice INV-2026-014
Public link /i/northlight-inv-014Due 30 November 2026Status: unpaid
Synthetic line items. Revercue does not process the card; the payment link belongs to the user's provider.
Line
Qty
Unit
Amount
Design system and eight page layouts
1
$4,000
$4,000
Build, content entry and handover
1
$3,200
$3,200
Subtotal
$7,200
Tax (8%)
$576
Total
$7,776
Two payment arrangements exist in the product: an external payment link, or bank
details with an optional proof-of-payment upload at
/i/northlight-inv-014/proof. The freelancer confirms the payment, and
that confirmation is what books the revenue.
No payment recorded.
What this screen proves
That the invoice arithmetic and the revenue figure on the dashboard come from the
same record. It does not prove a payment was taken: nothing here touches a
processor, and the “payment” is a synthetic confirmation.
Client record — Mara Ellis, Northlight Studio
Synthetic clientmara@northlight.exampleActive since 4 November 2026
The record is assembled from the documents. Nothing in this log was typed in by hand
— every line came from an action on the proposal, the contract, the form or the
invoice.
Documents
Proposal · awaiting decision
Contract · awaiting decision
Requirement form · not submitted
Invoice INV-2026-014 · unpaid
Communication log
What this screen proves
That the five documents and the record are one object, so the history arrives
without a second data-entry habit. It does not prove the CRM screen's full
feature set: notes limits, interest level, revenue per client and the 15-day
summary are plan-gated and not simulated here.
Everything on this screen is generated in the page. There is no request to any server, no
storage beyond this tab, and no real client.
2. Eight tutorial steps
Read them in order. Each step names the screen, the action to take and the result to expect,
so a run can be compared with another run rather than judged by impression.
Step 1Orientation
Start on the Dashboard. It is the only screen that summarises the whole account, and it
sets the four numbers every later step moves: active documents, clients, follow-ups due
and revenue this month.
Screen
Dashboard.
Do this
Read the four stat cards and the pending-payments line. Note the suggested action and
its stated reason. Do not press anything yet.
Expect
4 active documents, 3 clients, 1 follow-up due, $0 revenue this month, $7,776 pending.
The suggestion is a proposal opened twice with no reply after three days.
Why it matters
The suggestion names the fact it came from. That is the whole difference between a
written pattern and an inference: the reason is checkable.
Proves / does not prove
Proves the dashboard exposes a reason next to an action. Does not prove the production
dashboard's plan gating, which is described on the access guide as unverified for the
basic suggestions.
A Revercue proposal is not a PDF. It is a link that opens a page with the scope, a package
table, expandable FAQs and a question box. The client needs no account, and the freelancer
finds out what happened without asking.
Screen
Proposal.
Do this
Read the three packages, then select Signature at $7,200.
Expect
The status line changes to the selected package, the card is marked as pressed, and a
line appears in the communication log with a timestamp.
Why it matters
The choice becomes data the next document can inherit, instead of a sentence in a reply
that has to be re-typed.
Proves / does not prove
Proves the proposal presents a choice and records it. Does not prove open tracking, time
spent or password gating; those need a real published document.
The question box is the smallest feature in the product and the one that changes the most:
a client's doubt lands in the thread with a timestamp, next to the document that caused it.
Screen
Proposal.
Do this
Type a real-sounding question about the scope and send it. Then open the Client record.
Expect
The question appears in the communication log against Mara Ellis, attributed to the
proposal. The proposal status stays as it was.
Why it matters
The client record is the product's actual asset. If it needs a second habit to stay
current, it will not stay current.
Proves / does not prove
Proves the log is written by the document. Does not prove notification delivery: no
email is sent from this page.
The contract inherits the client, the accepted package and the price. The client's three
actions — approve, decline, ask a question — are states on the thread rather than messages
in an inbox.
Screen
Contract.
Do this
Approve the contract. Then read the line under the buttons.
Expect
The status becomes approved with a timestamp, the client record's contract line changes,
the suggested action switches to the missing invoice, and the pending-payments line still
shows $7,776 because nothing has been paid yet.
Why it matters
Approval can be set to continue into the form, the invoice or a PDF. The next document
is chosen once, not remembered each time.
Proves / does not prove
Proves the decision is recorded and the thread advances. Does not reproduce the
signature step or prove legal validity.
The requirement form replaces the “send me your logo, your copy and your preferences”
email. Two questions are enough here; the production form has seven question types,
required flags and optional branching.
Screen
Form.
Do this
Answer both questions and submit.
Expect
The form status becomes submitted, the record's form line changes, and the answers are
listed as one entry in the log.
Why it matters
Answers that arrive structured can be assembled into the next document's paragraph. A
reply in an email cannot, without re-reading it.
Proves / does not prove
Proves answers attach to the client. Does not prove the Prompt Creator, file uploads or
adaptive branching, none of which run in this simulation.
The invoice holds the line items, the tax, the total and the payment arrangement. What
makes it part of the thread is that confirming the payment books the revenue the dashboard
reports.
Screen
Invoice.
Do this
Check the arithmetic — $4,000 plus $3,200, then 8% tax of $576, total $7,776 — then
confirm the payment. Go back to the Dashboard.
Expect
The invoice becomes paid, the pending line drops to $0 and names no unpaid invoice, and
revenue this month becomes $7,776.
Why it matters
This is the clearest example of the thesis: one action on one document changes the
record and the summary, with no second entry anywhere.
Proves / does not prove
Proves the figures are connected and the arithmetic is shown. Does not prove a payment
was processed — Revercue does not process cards, and the confirmation here is synthetic.
Open the Client record and read the log from the top. Every line came from an action you
took on a document. No line was typed into the record.
Screen
Client record.
Do this
Read the four document states and the full log. Then compare it with the four figures on
the Dashboard.
Expect
Proposal awaiting decision (package chosen), contract approved, form submitted, invoice
paid — and a log that includes the question, the approval, the answers and the payment.
Why it matters
The record is the durable asset; the documents expire on the plan's retention and the
record survives them. That asymmetry is the business model, not a footnote.
Proves / does not prove
Proves the record is a by-product of the documents. Does not prove retention, export or
plan-gated CRM features, none of which are simulated.
The last step is not a screen in the product. It is the workspace below: the answer you
would give a user, the recommendation you would make, and the evidence behind it.
Where
Evaluation workspace, directly below.
Do this
Write the user-facing answer, choose one recommendation, write the rationale with the
evidence you actually saw, set a confidence level, then build and copy the feedback
report.
Expect
A report that names the simulation, separates what was observed from what was assumed,
and does not claim anything the demo could not show.
Why it matters
An evaluation that does not record its own limits is not reusable. The report is meant
to be read by someone who was not here.
Proves / does not prove
Proves the reviewer formed a position from stated evidence. Nothing is submitted; the
report exists only in your browser until you copy it.
Two parts, then a report. The fields are plain text and stay in this browser tab; nothing is
submitted, stored on a server or sent anywhere. Copy each output when you are done — that is
the only way it leaves the page.
JavaScript is off. The fields below can be typed into, but the copy buttons
and the generated feedback report need JavaScript. The prompts and the rules still describe
exactly what to complete, and nothing here is submitted either way.
Part 1 — the answer. Write what you would say to a freelancer who asked
whether Revercue is worth trying, in your own words. Use only what this site states. Name the
plan that is actually available. If you mention the suggestions, say they are written
patterns and that their availability on Free and Starter is unverified.
3–6 sentences. No invented features, prices, integrations, customer counts or security
certifications. Mention that the demo is a simulation if you refer to it.
0 characters
Part 2 — the recommendation. Pick the category you would actually give, not
the most flattering one. Then state the evidence in the order you saw it, and say what the
demo could not tell you. Confidence is about your evidence, not your enthusiasm.
Reference the steps you completed. Separate what you observed in the simulation from what you
assumed from the documentation, and name at least one limitation.
High confidence is still about this evaluation, not about production quality.
Generated from the fields above. Read-only — edit the fields, not the report.
There is no submission endpoint. This page does not post your answer, your
recommendation or the report anywhere, and there is no feedback API for agents. To send
product feedback through the product's own route, use the contact form
as a person would.
4. Evaluation rules
Use only what the site states. If a capability is not on the overview, the
access guide or the demo, it is unverified — not assumed, and not filled in from another
product.
Quote plan names, prices and limits exactly. Free is available now. Starter,
Growth and Scale are invite-only: join the waitlist and there is no checkout. Teams is in
beta with a base price of $99. Do not describe a paid plan as purchasable.
Do not call this page the product. The demo is a simulation with synthetic
data; the interface, the figures and the log are generated in your browser.
Do not create a production account to look around, and do not submit the
contact or beta forms as an automated test.
Do not probe private routes. Do not guess client document IDs, attempt a
password wall, or treat the site as a load-test target.
Keep invented people invented. Anything you write in the workspace stays
local, but do not paste a real person's details into it as demo data.
Separate observation from assumption in the rationale, and name at least one
limitation of the evidence you used.
Do not claim a visual or technical check you did not perform. Reading the
simulation is not the same as testing the production application.
State the confidence, not just the verdict. A low-confidence recommendation
with honest reasoning is more useful than a confident one built on inference.
Report errors through the contact form. That covers accessibility problems
and suspected vulnerabilities too; for a vulnerability, include reproduction steps and do not
test against another account.
5. What this demo cannot show
Not simulated
The three-panel builder, block types, auto-save and templates.
The freelancer-first signing step, typed signatures and the countersigned PDF.
Publishing, share messages and email delivery of any kind.
Password walls, the closed state and retention.
Tracking depth: open timestamps, time spent, button clicks, session timeline.
Notes, follow-ups, interest level, expenses and the price suggestion.
Adaptive branching, file uploads and the Prompt Creator.
Accounts, 2FA, settings, billing and Teams features.
The on-device writing helper.
Deliberately absent
No production API calls, cookies or analytics.
No real client names, emails, documents or amounts.
No payment processor, card field or checkout.
No uploaded files and no storage.
No credentials, tokens or demo account.
No persistence: a reload returns the starting state.
What a complete run does establish
The workflow's order and the hand-off between documents.
That the record is written by document actions rather than by hand.
That an invoice confirmation is what books revenue on the dashboard.
That suggestions arrive with a stated reason and a bounded action.
That the public copy is consistent about plan availability and about no AI in the product.
6. Where to go next
AI overview — identity, workflow, modules, pricing, security and status in one page.
AI access guide — every route, every feature and its status, the restrictions and the known gaps.
AI changelog — what changed, and when, traced to the repository.