X Twitter Thread Writer: Hook, Cadence, and Reply CTA

Prompt

Act as an X ghostwriter who ships threads people finish. You write for the 2026 timeline: a short first tweet, numbered posts, receipts, and no guru cadence.

Inputs:
- Topic or claim: [Claim]
- Audience: [Audience]
- Goal: [Follows / replies / clicks / waitlist]
- Proof I can use: [Proof]
- Voice: [Voice]
- Length: [5-7 tweets / 8-12 tweets]
- Link I may include: [Link or none]
- Things I will not say: [Banned claims]
- Character comfort: [Punchy under 200 / mix / longer 240-280]

Generate:
1. Positioning: One sentence on the unique angle vs. the generic thread on this topic.
2. Hook options (5): First tweets under 240 characters. Mix: specific number, confession, pattern interrupt, named mistake, question. Mark one as recommended.
3. Full thread: Write the complete numbered thread at the requested length. Tweet 1 is the recommended hook. Each later tweet is one idea. Last tweet is a reply prompt, not "agree?" or "RT this." If a link is allowed, put it only in the last or second-to-last tweet. Number tweets 1/N style.
4. Receipts map: For every number, story, or "I did X," point at Proof I can use. If Proof cannot support it, cut the claim or label it as opinion.
5. First reply: The reply I should post under my own thread (caveat, extra example, or resource). Keep it under 200 characters.
6. Cut list: Three tweets I would delete if the thread must shrink by 25 percent, and why.
7. Mute check: List any line that sounds like a growth account. Rewrite it.

Constraints:
- Ban: "A thread:", "Let's dive in", "Unpopular opinion" for popular takes, "You are not lazy you are...", emoji walls, "RT if", "I'll wait", fake screenshots, hashtag dumps.
- Do not invent metrics, studies, follower counts, or screenshots. Use only Proof I can use.
- Sound like a person who did the work, not a thread bro.
- Honor Things I will not say.
- At most two niche tags, only if they help discovery, and only on the last tweet.
- No emojis unless Voice asks for them.

Instructions

Replace every [bracket] with your details before running. Works on ChatGPT, Claude, and Gemini.

Examples

Example Input

Topic or claim: shipping a weekly public changelog beat a fancy launch for our B2B tool
Audience: indie SaaS founders and product-led growth operators
Goal: follows
Proof I can use: we posted a Friday changelog for 16 weeks; waitlist-to-paid rose from 4% to 7% in that window; support tickets about "what's new" dropped; we did not run ads; team of 3
Voice: Direct, slightly dry, founder-to-founder
Length: 8-12 tweets
Link I may include: none
Things I will not say: overnight success, we 10x'd, growth hacking, guru frameworks
Character comfort: Punchy under 200

Example Output

1. Positioning
The angle is operational: a boring Friday changelog as a sales motion, not "build in public as a lifestyle."

2. Hook options (5)
A. We skipped the launch video. 16 Friday changelogs later, waitlist-to-paid went 4% to 7%. (recommended)
B. The most expensive page we never shipped was the big launch.
C. I thought we needed a launch. What we needed was a Friday habit the three of us would not skip.
D. Stop hiding the product behind a "coming soon" and then wondering why no one trusts the demo.
E. What if the launch is just the 16th changelog with a nicer headline?

3. Full thread
1/11 We skipped the launch video. 16 Friday changelogs later, waitlist-to-paid went 4% to 7%.

2/11 We are a team of 3. No ads. The changelog is a Notion page dumped into an email and an X post. Ugly on purpose.

3/11 The old motion was: disappear for six weeks, ship a "big" feature, write a thread, watch a spike, watch it die.

4/11 Buyers in B2B do not need a moment. They need proof you are still alive, still shipping, and still honest about what broke.

5/11 Format we reused: 3 bullets of shipped, 1 bullet of what slipped, 1 screenshot if we had one. Same day every week. Friday 10am.

6/11 The 4% to 7% move showed up in the waitlist we already had. I am not claiming a category shift. I am claiming fewer silent drop-offs.

7/11 Side effect we did not plan: support stopped answering "did you ship X yet?" The changelog became the answer. That is time a 3-person team actually feels.

8/11 What we refused: a teaser week, a waitlist fake countdown, a "we 10x'd" recap. If a week was thin, the post was thin. People trusted the thin weeks.

9/11 If you want to steal this, steal the cadence, not the template. Pick a day. Name an owner. Publish even when the week is embarrassing.

10/11 The launch energy went into week 16: we summarized the 16 posts as "here is the product now." Same facts. Better table of contents.

11/11 If you already ship weekly and still stall, say what you publish and where it dies (email, X, in-app). I will tell you which one we would cut first.

4. Receipts map
- 16 weeks of Friday changelogs: Proof.
- Waitlist-to-paid 4% to 7%: Proof. Opinion: "fewer silent drop-offs" is a reading of that number, labeled as such.
- No ads, team of 3: Proof.
- Support tickets about "what's new" dropped: Proof (no count given, so the thread stays qualitative).
- Not claimed: revenue, follower growth, "this works for every SaaS."

5. First reply
Week 4 was two bugfixes and a copy change. We still posted. That post got more "ok these people are real" replies than week 12's actual feature.

6. Cut list
- Tweet 3 (old motion): redundant once tweet 4 lands.
- Tweet 10 (week 16 recap): nice, not required for the habit thesis.
- Tweet 8 second sentence: repeats the honesty point in tweet 4.

7. Mute check
"If you want to steal this, steal the cadence" is close to coach-speak. Leave it, but the next line (owner + embarrassing week) keeps it operational. No "build in public" slogan used.