Shelf
See how it works · Cold-to-Reply

How Cold-to-Reply works

Your expertise, faster.

Here's the full flow, end to end. A sales rep targets a Series B founder who announced a funding round. The form captures the offer, the sender, the prospect, the writing style, and the sequence shape. The AI returns a 4-touch sequence calibrated to the funding-raise signal, plus reply handlers for the top 5 reply types.

The actual tool runs locally and works the way you'll see below. This page is the preview.

Step 01

The tool opens with five empty sections

Double-click Cold-to-Reply.html from your downloads. It opens in any browser. No install, no account, no API key.

file:///Cold-to-Reply.html
Shelf HQ v1.3
Section 1 of 5

What you're selling

The offer plus the ICP. Specific is better than general.

e.g. A team-analytics dashboard...
Section 2 of 5

About you (the sender)

Your name, role, anything credibility-relevant...
Sections 3-5

Who you're reaching out to. Show me how you write. Shape the sequence.

The prospect (with signal), 1-3 writing samples, sequence shape.

Pastes a ready-to-go prompt for ChatGPT, Claude, or Gemini.
Step 02

Describe the offer and the sender

The buyer here is a sales rep at an analytics SaaS reaching out to a Series B founder. The first two sections cover what's being offered and who's sending it. The more credibility-relevant detail in Section 2, the more the sequence sounds like a real person reaching out.

Section 1 of 5

What you're selling

The offer
A team-analytics dashboard for SaaS companies that combines Stripe, Hubspot, and Mixpanel into one weekly review.
Your ICP
Series A through C SaaS, 20-200 employees, founder still in the room for top-10 customer reviews.
Section 2 of 5

About you (the sender)

Name and role
Marcus Chen, founding sales lead at Acme Analytics
Credibility / context
I was a PM at two prior B2B SaaS companies before this. I know how often weekly customer reviews get rescheduled because the data's a mess. We built Acme because I personally got tired of spending Sunday nights stitching together three dashboards before Monday's pipeline call.
Step 03

The prospect, the signal, your writing samples, the sequence shape

Section 3 names the specific prospect and the signal that triggered the outreach (just-raised funding, new hire, recent content, warm intro, or cold-cold). Section 4 is the writing-style match step. Section 5 shapes the sequence: signal type, touch count, optional reply handlers, optional LinkedIn variant.

Section 3 of 5

Who you're reaching out to

Prospect
Sarah Patel, founder/CEO of Northwind, a Series B B2B SaaS. Announced a $25M round last week.
The signal
Just-raised funding
New hire / role change
Recent content
Warm intro
Cold-cold (no signal)
Why now (the personal angle)
Post-Series-B, the weekly customer review is about to become impossible without tooling. She's about to scale ops; the data sprawl gets worse before it gets better.
Section 4 of 5

Show me how you write

Paste 1-3 short samples. The AI calibrates the sequence to your sentence rhythm.

Sample 1
Spent yesterday on a customer call where the founder pulled up three different dashboards to explain why churn dropped. He couldn't tell which one was right. By the end of the call he'd promised himself he'd "finally fix the data thing." That promise has a shelf life of about two days.
Sample 2
Most cold emails I get could be sent to anyone. The good ones can only be sent to me. Three sentences in, you can tell. I try to write the second kind.
Section 5 of 5

Shape the sequence

Touch count
4 touches over 3 weeks
Include reply handlers (top 5 reply types)
Include LinkedIn variant of touch 1
Pastes a ready-to-go prompt for ChatGPT, Claude, or Gemini.
Step 04

Click Build. The full prompt assembles.

One click. The tool stitches your offer, your sender context, the prospect-with-signal, your writing samples, and the sequence shape into one paste-ready prompt.

Your prompt is ready

You are writing a multi-touch cold outreach sequence on behalf of a sender. Match the sender's writing style from the samples below. Calibrate the hook to the prospect's specific signal. THE OFFER: A team-analytics dashboard for SaaS companies that combines Stripe, Hubspot, and Mixpanel into one weekly review. ICP: Series A through C SaaS, 20-200 employees. ABOUT THE SENDER: Marcus Chen, founding sales lead at Acme Analytics. Was a PM at two prior B2B SaaS before this... THE PROSPECT: Sarah Patel, founder/CEO of Northwind, a Series B B2B SaaS that announced a $25M round last week. THE SIGNAL: just-raised funding WHY NOW: Post-Series-B, the weekly customer review is about to become impossible without tooling. ===== THE SENDER'S WRITING STYLE (FOR PATTERN, NEVER FOR COPYING) ===== [2 writing samples embedded] The samples above are STYLE references. Match register, sentence rhythm, opener pattern. Do NOT copy exact phrases. ===== SEQUENCE SHAPE ===== - 4 touches over 3 weeks - Touch 1: cold opener acknowledging the funding signal - Touch 2 (5 days later): a specific data point relevant to founder - Touch 3 (7 days later): one short proof point, no ask - Touch 4 (14 days later): graceful goodbye - Plus: LinkedIn variant of touch 1 - Plus: 5 reply handlers (Not now / Send more info / Already have a tool / Wrong contact / Not interested) ===== HUMANIZE PASS (REQUIRED) ===== After the sequence, scrub for AI-tells [full banned-word list embedded] ===== OUTPUT RULES (STRICT) ===== - Output every touch and every handler in one response, clearly delimited - No preamble. No clarifying questions. No meta-commentary.
Step 05

Paste into your AI. The full sequence comes back.

Open a fresh chat in ChatGPT, Claude, or Gemini. Paste. Hit send. The 4-touch sequence comes back with each touch labeled, plus reply handlers for the top 5 reply types and an optional LinkedIn variant of touch 1.

Claude Fresh chat. Sequence received.
Touch 01 Day 0 · email

Subject: post-Series-B data sprawl

Sarah, congrats on the round.

The thing nobody warns you about post-Series-B: the weekly customer review goes from "an hour" to "the worst meeting on your calendar." The data sprawl gets worse before it gets better. Stripe says one number. Hubspot says another. Mixpanel says something nobody can interpret.

We built Acme because I personally got tired of spending Sunday nights stitching three dashboards before Monday's pipeline call. We work with Series A through C SaaS, 20-200 people. The whole product is one weekly review, three sources, one number per metric.

If post-round ops scaling is on your radar, worth a 20-minute call?

Marcus

Touch 02 Day 5 · email

Subject: one data point

Sarah,

One thing we've seen at three Series B SaaS we work with: the metric that quietly breaks first is "active accounts in the last 14 days." Stripe doesn't see it. Hubspot doesn't see it. Mixpanel sees it but nobody trusts the cohorting.

If you want, I can send you the queries we use to triangulate it across all three. No call needed. Reply yes and I'll send Friday.

Marcus

Touch 03 Day 12 · email · short

Three-line proof point about a Northwind-comparable customer who cut their weekly review prep from 3 hours to 40 minutes. No ask in this touch. Only proof.

Touch 04 Day 26 · email · graceful exit

Two lines. Acknowledges the silence is fine. Leaves the door open for a future inbound. No re-pitch.

Reply handlers (5 included)
  • Not now: Calibrate timing question. Offer a short async option.
  • Send more info: Two-paragraph product pager, not a deck.
  • Already have a tool: Acknowledge. Ask one specific differentiation question.
  • Wrong contact: Thank, ask for the right one, no follow-up to Sarah.
  • Not interested: Acknowledge. Offer a permission-to-revisit-in-6-months line.

That's the whole flow.

From a prospect name and a funding-round signal to a 4-touch sequence with reply handlers, in your writing style, in under 5 minutes of your time.

The 15% that matters (the recipient's specific context, the personal reference no AI knows about, your read of their actual signal) is yours.

$18
One-time. Lifetime updates. 7-day refund if the product doesn't work as described.
Powered by Gumroad
Walkthrough captured on v1.3.