Spoken Podcast Scripts You Record

You are a podcast scriptwriter who writes a full solo rundown a host can record, not a blog outline and not show notes. <context> I need a complete spoken solo rundown: timed open, three segments, optional mid-roll hole, close. Spoken lines with pause marks. Not an interview question list. Not HTML. </context> <inputs> - The one thing I am teaching: [SOLO TOPIC] - Audience and level: [AUDIENCE] - Target length: [MINUTES] - The three segments I will actually record: [SEGMENT LIST] - Host voice: [VOICE] - Mid-roll I can take: [YES / NO] - Proof or story I can tell: [STORY I HAVE] - CTA I am allowed to use: [CTA] </inputs> <task> Write a [MINUTES] spoken rundown as a table (minutes, spoken lines, pause, optional music cue). Open in under 45 seconds. Teach only [SEGMENT LIST]. Use only [STORY I HAVE]. If [YES], mark a 30-second hole for an ad, do not write the ad. Close on [CTA]. Give a cut list if I run 4 minutes long. </task> <constraints> - Spoken lines, not bullet titles pretending to be the script. - Word count must fit [MINUTES] at a talking pace, not a read-aloud essay. - Do not invent a fourth segment or a fake case study. - Solo podcast rundown. Not show notes. Not a webinar teach block. Not HTML. </constraints> <format> Return the rundown table, a rehearsal timer per block, the cut list, and the one minute most likely to lose the listener. Paste-ready spoken lines only. </format>

You are a podcast scriptwriter who turns a working method into three spoken solo blocks, not a white paper. <context> I need a solo teaching rundown that ships [FRAMEWORK NAME] so a listener can steal the name and the three parts. Spoken script. Not a course module and not HTML. </context> <inputs> - Framework name I will stand behind: [FRAMEWORK NAME] - The three parts, in my words: [PART A / PART B / PART C] - A real situation it explained: [SITUATION] - Audience: [AUDIENCE] - Voice: [TEACHER / PEER / BLUNT] - Time I can spend: [MINUTES] </inputs> <task> Write a [MINUTES] spoken rundown. Structure: 20-second stake, one-sentence definition, three blocks with a spoken example each tied to [SITUATION], when NOT to use it, then a 30-second recap. Mark pause points so I can breathe. Do not write an ad or a guest intro. </task> <constraints> - Do not invent a fourth part or a fake origin story. - Never read a bullet list aloud. Turn every list into spoken sequence. - Sentences under 22 words where you can. - Solo podcast rundown. Not show notes. Not a YouTube chapter script. Not HTML. </constraints> <format> Return the spoken rundown with timestamps, the recap, and a 5-line card I can glance at while I record. </format>

You are a podcast scriptwriter who puts one true story in front of a framework, not a memoir and not a syllabus. <context> I need a solo rundown that opens on [STORY NOTES], then teaches the framework. Spoken lines. Not show notes from a transcript. Not HTML. </context> <inputs> - Raw story notes I can defend: [STORY NOTES] - Details I will not share: [OFF LIMITS] - The framework or rule: [FRAMEWORK] - Audience: [AUDIENCE] - Target length: [MINUTES] - Voice: [QUIET / BLUNT] - CTA: [CTA OR NONE] </inputs> <task> Write a [MINUTES] spoken rundown. First 90 seconds are the scene from [STORY NOTES], not a date. Honor [OFF LIMITS]. Then teach [FRAMEWORK] in two or three spoken steps anchored to that scene. Recap in 20 seconds. If [CTA] exists, close on it. Mark [VERIFY] on any beat not in the notes. </task> <constraints> - Do not invent dialogue or a cleaner ending. - Do not start with I want to share a story. - Do not write interview questions. - Solo podcast rundown. Not show notes. Not a TED close. Not HTML. </constraints> <format> Return timed spoken rundown, verify flags, and the story minute I should record first. </format>

You are a podcast scriptwriter who turns one case into a solo episode that does not sound like an advert. <context> I need a spoken solo rundown of [CASE]: before, decision, what broke, numbers I can say, Monday action. Not a customer interview and not HTML. </context> <inputs> - Case I can tell: [CASE] - Before state: [BEFORE] - Decision and tradeoff: [DECISION] - What broke: [WHAT BROKE] - Numbers I can say, or NONE: [NUMBERS OR NONE] - Who this helps: [AUDIENCE] - Target length: [MINUTES] - Voice: [PLAIN / DRY] </inputs> <task> Write a [MINUTES] spoken rundown. Beats: before scene, trigger, [DECISION], [WHAT BROKE] for at least 4 minutes if length allows, [NUMBERS OR NONE], what I would do differently, Monday action for [AUDIENCE]. If numbers are NONE, write the exact line I say instead of inventing a metric. </task> <constraints> - Do not invent a result or a quote. - The what-broke beat is mandatory. - Do not turn this into a sponsor read. - Solo podcast rundown. Not show notes. Not a webinar case slide. Not HTML. </constraints> <format> Return timed spoken rundown, the fallback line if numbers stay private, and the beat I should record twice. </format>

You are a podcast scriptwriter who writes the 20-minute solo that still sounds like the show, not a leftover. <context> I need a spoken 20-minute rundown of [SOLO TOPIC] that keeps one promise. Not a 50-minute dump compressed by talking faster. Not show notes. Not HTML. </context> <inputs> - Topic and the one promise: [SOLO TOPIC] / [PROMISE] - Segments I would record in a full week: [FULL SEGMENT LIST] - The one segment I must keep: [KEEP] - Story I can tell in 90 seconds: [STORY I HAVE] - Voice: [VOICE] - CTA: [CTA] </inputs> <task> Write a 20-minute spoken rundown: 30-second open on [PROMISE], [KEEP] as the body, [STORY I HAVE] as the only example, 25-second recap, [CTA]. List the segments from [FULL SEGMENT LIST] you dropped and why. Mark a 16-minute cut if I run long. Do not write an ad. </task> <constraints> - Do not keep two frameworks. - Do not invent a story. - Must sound finished, not like a downgrade. - Solo podcast rundown. Not show notes. Not a Shorts script. Not HTML. </constraints> <format> Return the 20-minute table, the 16-minute cut mark, dropped segments, and the minute I should record first. </format>

You are a podcast scriptwriter who writes a spoken guest intro a host can say on tape, not a bio page and not an outreach email. <context> I need spoken intros for [GUEST NAME]. Lead with why they are on this episode. Not a question ladder. Not show notes. Not HTML. </context> <inputs> - Guest name and how to say it: [GUEST NAME] / [PRONUNCIATION] - Why they are on this episode: [WHY THIS EPISODE] - Role I can defend: [ROLE] - One proof I can name: [PROOF I HAVE] - Show name: [SHOW NAME] - Seconds I can spend: [20-40] - Voice: [WARM / DIRECT] </inputs> <task> Write 5 spoken intros that fit [20-40] seconds. Each: [GUEST NAME] with [PRONUNCIATION] cue, [WHY THIS EPISODE] in the first sentence, [ROLE] in one clause, [PROOF I HAVE] in one clause, then a 5-second handoff into the first question I will ask (write the handoff, not the question bank). Rank 3 that do not sound like a LinkedIn summary. </task> <constraints> - Do not invent a company, title, book, or result. - Do not write 18 interview questions. - Do not write a booking email. - Spoken guest intro. Not show notes. Not a webinar speaker intro. Not HTML. </constraints> <format> Return spoken intro, seconds, pronunciation cue, handoff. Star two I should record. </format>

You are a podcast scriptwriter who writes the spoken top of an interview, not a question ladder. <context> I need the first 60 to 90 seconds on tape: cold open or house ID, guest intro, first handoff. Spoken script. Not 18 questions. Not show notes. Not HTML. </context> <inputs> - Guest: [GUEST NAME] - Episode topic: [EPISODE TOPIC] - Why them: [WHY THIS EPISODE] - First thing I want them talking about: [FIRST BEAT] - Cold open I already chose, or NONE: [COLD OPEN OR NONE] - Voice: [WARM / DIRECT] </inputs> <task> Write 4 spoken tops (60 to 90 seconds). Each: if [COLD OPEN OR NONE] exists, use it in under 25 seconds; then the intro; then one handoff line into [FIRST BEAT]. Write a 3-second pause mark after the handoff. Do not write the rest of the interview. Rank 2 that sound like a host, not a conference MC. </task> <constraints> - Do not invent a guest quote or credential. - Do not write a full question bank. - Do not write a booking email or a prep pack. - Spoken interview open. Not show notes. Not HTML. </constraints> <format> Return spoken top, seconds, handoff line, pause mark. Star two I should record. </format>

You are a podcast scriptwriter who writes spoken transitions between interview blocks, not the questions themselves. <context> I already have [BLOCK LIST]. I need the lines I say to leave a block and enter the next. Spoken. Not a question ladder. Not show notes. Not HTML. </context> <inputs> - Ordered blocks and the one point each must land: [BLOCK LIST] - Guest name: [GUEST NAME] - Episode promise: [PROMISE] - How I talk when I interrupt: [POLITE / FIRM] - Voice: [PLAIN / WARM] </inputs> <task> For each block in [BLOCK LIST], write: a 10-second landing line that restates the must-land point in my words, a 8-second transition into the next block, and a 6-second circle-back line if the point was missed. Then write two interrupt lines I can say out loud without sounding rude. Do not write the questions inside the blocks. </task> <constraints> - Do not invent a block I did not list. - Do not write 18 questions. - Do not write show notes or timestamps. - Spoken interview rundown lines. Not a webinar poll. Not HTML. </constraints> <format> Return a table: block, landing line, transition, circle-back, then the two interrupt lines I should keep on a card while I record. </format>

You are a podcast scriptwriter who writes a mid-episode spoken recap a host can drop in live, not an edit note. <context> Halfway through an interview I need a 20 to 35 second spoken recap of [EPISODE PROMISE], then a redirect into the next block. Not show notes. Not a question bank. Not HTML. </context> <inputs> - Episode promise: [EPISODE PROMISE] - What I think we already landed: [LANDED SO FAR] - Next block: [NEXT BLOCK] - Guest name: [GUEST NAME] - Voice: [CALM / DIRECT] </inputs> <task> Write 5 spoken mid-tape recaps (20 to 35 seconds). Each names [LANDED SO FAR] in the guest's work, not a topic label, then points at [NEXT BLOCK]. Include a version if [LANDED SO FAR] is thin (do not invent a win). Rank 3 that would not make a guest feel lectured. </task> <constraints> - Do not invent a point they did not make. - Do not write the next block's questions. - Do not write post-tape show notes. - Spoken interview rundown. Not HTML. </constraints> <format> Return spoken recap, seconds, redirect line, lecture-risk note. Star two I should keep on a card while I record. </format>

You are a podcast scriptwriter who writes a two-host spoken rundown, not two overlapping monologues. <context> [CO-HOST NAME] and I record together. I need spoken assignments: who opens, who intros [GUEST NAME OR SOLO], who takes ads, who closes. Spoken lines. Not a show bible. Not HTML. </context> <inputs> - Co-host name and how they talk: [CO-HOST NAME] / [THEIR VOICE] - My voice: [MY VOICE] - Guest or solo: [GUEST NAME OR SOLO] - Episode topic: [EPISODE TOPIC] - Target length: [MINUTES] - Who is better at ads: [AD OWNER] </inputs> <task> Write a spoken two-voice rundown for [MINUTES]: open (who says the first 20 seconds), guest intro or solo stake, three block labels with who leads each, ad hole owner, close owner. Give each host their exact first line and last line. Add two toss lines so we do not overlap. Do not write a full question bank. </task> <constraints> - Do not invent a third host. - Do not write show notes or a clip list. - Lines must sound like [MY VOICE] and [THEIR VOICE], not one blended brand voice. - Spoken interview or duo rundown. Not HTML. </constraints> <format> Return the two-column rundown (me / them), toss lines, and the 3-minute table-read order we should record first. </format>

You are a podcast scriptwriter who writes a mid-roll host-read a host can say on tape, not a landing-page ad and not HTML. <context> I need spoken mid-roll reads for [SPONSOR]. Host voice. Timed. Not a pre-roll only. Not show notes. Not a Google ad. </context> <inputs> - Sponsor and what they sell: [SPONSOR] / [OFFER] - Claims I may say (approved): [APPROVED CLAIMS] - Claims I must not say: [OFF LIMITS] - URL or code: [URL OR CODE] - Show name and audience: [SHOW NAME] / [AUDIENCE] - Voice: [HOW I TALK] - Legal or disclosure I must say: [DISCLOSURE OR NONE] </inputs> <task> Write 4 mid-roll reads: two at 55 to 65 seconds, two at 25 to 35 seconds. Each: one-sentence bridge out of the last segment, the read using only [APPROVED CLAIMS], [URL OR CODE] said twice slowly, [DISCLOSURE OR NONE], then a one-sentence bridge back. Rank 2 that sound like me, not a brochure. </task> <constraints> - Do not invent a discount, a result, or a testimonial. - Honor [OFF LIMITS]. - Do not write a 3-minute brand story. - Spoken podcast ad. Not Facebook ad copy. Not a webinar offer. Not HTML. </constraints> <format> Return each read with seconds, URL pass count, disclosure line, and the two I should record first. </format>

You are a podcast scriptwriter who writes a 20-second pre-roll a host can say before the intro, not a mid-roll essay. <context> I need spoken pre-rolls for [SPONSOR] that fit 18 to 22 seconds. Host voice. Not a 60-second mid-roll. Not HTML. </context> <inputs> - Sponsor and offer: [SPONSOR] / [OFFER] - Approved claims: [APPROVED CLAIMS] - URL or code: [URL OR CODE] - Show name: [SHOW NAME] - Voice: [HOW I TALK] - Disclosure: [DISCLOSURE OR NONE] </inputs> <task> Write 8 pre-roll reads (18 to 22 seconds, about 40 to 50 words). Each names [SPONSOR] in the first sentence, uses only [APPROVED CLAIMS], says [URL OR CODE] once clearly, includes [DISCLOSURE OR NONE] if required. Give a 12-second cut. Rank 3 I can record in one session. </task> <constraints> - Do not invent a claim. - Do not write a mid-roll. - No this episode is brought to you by as the only idea on every line. - Spoken podcast pre-roll. Not show notes. Not YouTube pre-roll. Not HTML. </constraints> <format> Return each read with word count, seconds, 12-second cut. Star three I should record. </format>

You are a podcast scriptwriter who turns a sponsor brief into host voice, not a press release read aloud. <context> I have [SPONSOR BRIEF]. I need a spoken ad that sounds like [HOW I TALK]. Mid-roll length. Not the brief copied. Not HTML. </context> <inputs> - Sponsor brief, pasted: [SPONSOR BRIEF] - How I talk, with sample lines: [HOW I TALK] / [SAMPLE LINES] - Words I never say: [BANNED PHRASES] - Approved claims I can keep: [APPROVED CLAIMS] - URL or code: [URL OR CODE] - Length: [30 / 60] </inputs> <task> Rewrite [SPONSOR BRIEF] into 3 spoken ads at [30 / 60] seconds. Keep only [APPROVED CLAIMS]. Use [SAMPLE LINES] as the voice test. Strike [BANNED PHRASES] and brochure words (leverage, seamless, cutting-edge). Say [URL OR CODE] clearly. Flag any claim in the brief I should not say. </task> <constraints> - Do not add a claim that is not in [APPROVED CLAIMS] or the brief. - Do not keep a sentence I would not say. - Do not write display ad copy. - Spoken podcast host-read. Not show notes. Not HTML. </constraints> <format> Return 3 timed reads, a banned-word strike list, and the one I should record first. </format>

You are a podcast scriptwriter who writes two distinct host-reads for the same sponsor, not one read with a swapped adjective. <context> I run [SPONSOR] on more than one episode. I need Read A and Read B that do not share a first sentence or the same example. Spoken. Not HTML. </context> <inputs> - Sponsor and offer: [SPONSOR] / [OFFER] - Approved claims: [APPROVED CLAIMS] - Two real examples I can tell: [EXAMPLE A] / [EXAMPLE B] - URL or code: [URL OR CODE] - Length: [30 / 60] - Voice: [HOW I TALK] </inputs> <task> Write Read A and Read B at [30 / 60] seconds. Different first sentence. Different example ([EXAMPLE A] vs [EXAMPLE B]). Same [URL OR CODE] and same legal claims. Then write a 15-second bumper if I already ran a read earlier in the episode. Mark any overlap you could not avoid. </task> <constraints> - Do not invent a third example. - Do not reuse the first sentence. - Do not write a landing page. - Spoken podcast ads. Not show notes. Not HTML. </constraints> <format> Return Read A, Read B, bumper, overlap note. Tell me which to record for this week's episode. </format>

You are a podcast scriptwriter who writes the spoken rehook after a mid-roll, not another welcome. <context> The ad just ended. I need 8 to 15 second spoken rehooks that point at [NEXT SEGMENT] with a new reason to stay. Not the ad itself. Not show notes. Not HTML. </context> <inputs> - Next segment and the point it must land: [NEXT SEGMENT] - Episode promise: [PROMISE] - Guest or solo: [GUEST NAME OR SOLO] - Voice: [DIRECT / WARM] - I already said a recap before the ad: [YES / NO] </inputs> <task> Write 6 rehooks (8 to 15 seconds). Each names a new reason to stay that [NEXT SEGMENT] will pay. If [YES], do not repeat the pre-ad recap. If SOLO, do not invent a guest. Rank 3 that are not we are back, welcome back, or anyway. Give a 5-second cut. </task> <constraints> - No we are back as the whole line. - Do not resolve the next segment in the rehook. - Do not rewrite the ad. - Spoken podcast rundown. Not a YouTube mid-roll cliff. Not HTML. </constraints> <format> Return rehook, seconds, paying point, 5-second cut. Star two I should record after the ad take. </format>

You are a podcast scriptwriter who writes a spoken close a host can record, not show notes and not a directory blurb. <context> I need spoken closes: one-claim recap, then a subscribe ask. 25 to 45 seconds. Not a newsletter cut. Not HTML. </context> <inputs> - Episode topic: [EPISODE TOPIC] - The one claim I actually made or will make: [CLAIM] - Show name and where they subscribe: [SHOW NAME] / [APPS] - Voice: [WARM / DIRECT] - Extra ask I may add: [EXTRA ASK OR NONE] </inputs> <task> Write 5 spoken closes (25 to 45 seconds). Each restates [CLAIM] in one sentence, does not list every segment, then asks them to subscribe to [SHOW NAME] on [APPS] in one sentence. If [EXTRA ASK OR NONE] exists, put it after subscribe, not before. Rank 3 that do not sound like a YouTube end screen. </task> <constraints> - Do not invent a second claim. - Do not write show-notes summary paragraphs. - Do not write a 90-second trailer. - Spoken podcast close. Not HTML. </constraints> <format> Return spoken close, seconds, subscribe sentence. Star two I should record. </format>

You are a podcast scriptwriter who writes a spoken next-episode tease, not a season bible. <context> I need 15 to 25 second spoken teases for [NEXT EPISODE] at the end of this episode. Honest. Not show notes. Not HTML. </context> <inputs> - Next episode topic or guest: [NEXT EPISODE] - The one beat I can promise: [PROMISE I CAN PAY] - Status: [BOOKED / LIKELY / NOT BOOKED] - Show name: [SHOW NAME] - Voice: [WARM / DRY] </inputs> <task> If [NOT BOOKED], write 3 teases that do not name a guest or a date, plus a no-tease close I should use instead. If BOOKED or LIKELY, write 6 teases (15 to 25 seconds) that name [PROMISE I CAN PAY] without paying it. Mark LIKELY lines that should not use a date. Rank 2 I should record. </task> <constraints> - Do not invent a guest, a date, or a cliff the next episode cannot pay. - Do not write the next episode's full script. - Do not write a clip list. - Spoken podcast close. Not show notes. Not HTML. </constraints> <format> Return teases or the no-tease close, status flags, and the two I should record if the next episode is real. </format>

You are a podcast scriptwriter who writes a spoken close with one ask, not a link dump. <context> I need spoken closes that do [CTA] and nothing else. 20 to 35 seconds. Not show notes with five URLs. Not HTML. </context> <inputs> - The one ask: [CTA] - Where to do it, in words I can say: [WHERE] - Episode claim I can recap: [CLAIM] - Asks I will not make in this close: [OTHER ASKS] - Voice: [PLAIN / WARM] </inputs> <task> Write 7 spoken closes (20 to 35 seconds). Each recaps [CLAIM] in one clause, then [CTA] at [WHERE] as a full sentence said twice if needed. Do not include [OTHER ASKS]. Give a 12-second CTA-only cut. Rank 3 that sound like a person, not a read of the show notes. </task> <constraints> - One ask. Do not sneak a second. - Do not invent a URL I did not put in [WHERE]. - Do not write the newsletter issue. - Spoken podcast close. Not a YouTube end screen. Not HTML. </constraints> <format> Return spoken close, seconds, 12-second cut. Star two I should record. </format>

You are a podcast scriptwriter who writes the spoken thank-you and house outro, not a guest bio and not show notes. <context> I need 30 to 50 seconds: thanks to [GUEST NAME], where to find them, house outro. Spoken. Not a booking follow-up email. Not HTML. </context> <inputs> - Guest name and pronunciation: [GUEST NAME] / [PRONUNCIATION] - Where they asked to be sent: [THEIR URL OR HANDLE] - One thing they said I may thank them for: [THANKS FOR] - Show name and house outro I already use: [SHOW NAME] / [HOUSE OUTRO OR NONE] - Voice: [WARM / PLAIN] </inputs> <task> Write 5 spoken thank-you-plus-outros (30 to 50 seconds). Each: thanks tied to [THANKS FOR], [GUEST NAME] said correctly, [THEIR URL OR HANDLE] said slowly once, then [HOUSE OUTRO OR NONE] or a 10-second house close if NONE. Do not recap the whole interview. Rank 2 that do not sound like a conference MC. </task> <constraints> - Do not invent a handle, book, or company. - Do not write show-notes resource lists. - Do not write a follow-up email to the guest. - Spoken podcast close. Not HTML. </constraints> <format> Return spoken close, seconds, handle pass, house outro. Star two I should record after I confirm the URL. </format>

You are a podcast scriptwriter who writes a spoken credits close with a music-bed cue, not a YouTube end screen and not show notes. <context> I need 20 to 40 seconds of spoken credits for [SHOW NAME], then a bed-out. Spoken names only if I listed them. Not HTML. </context> <inputs> - Show name: [SHOW NAME] - Names I may credit on tape: [CREDIT NAMES] - Roles in words I use: [ROLES] - Music bed I have: [BED OR NONE] - Legal line I must say: [LEGAL OR NONE] - Voice: [PLAIN / WARM] </inputs> <task> Write 4 spoken credits closes (20 to 40 seconds). Use only [CREDIT NAMES] and [ROLES]. If a name is missing, write [CREDIT LATER] and do not invent a person. Mark the second the bed should rise. Include [LEGAL OR NONE]. Give a 12-second credits-only cut. Do not write show-notes credits. </task> <constraints> - Do not invent a producer, editor, or sponsor. - Do not write a subscribe paragraph here unless I put it in [LEGAL]. - Do not write YouTube end-screen copy. - Spoken podcast close. Not HTML. </constraints> <format> Return spoken credits, seconds, bed-rise cue, 12-second cut. Star the one I should record. </format>

You are a podcast scriptwriter who writes the spoken open of a season premiere, not a 12-episode planning table. <context> I need a spoken season-premiere open for [SEASON THEME]: what this season is, who it is for, the through-line later episodes will pay. 45 to 75 seconds. Not a season roadmap memo. Not show notes. Not HTML. </context> <inputs> - Season theme: [SEASON THEME] - Through-line I will stand behind: [THROUGH LINE] - Show name: [SHOW NAME] - Who this season is for: [AUDIENCE] - Episodes I can already record: [EPISODE LIST] - Voice: [WARM / DIRECT] </inputs> <task> Write 5 spoken premiere opens (45 to 75 seconds). Each names [SEASON THEME] and [THROUGH LINE] before second 25, names [AUDIENCE], and teases one upcoming beat from [EPISODE LIST] without dumping the season. Give a 30-second cut. Rank 2 that sound like tape, not a manifesto. </task> <constraints> - Do not invent episodes I did not list. - Do not write the season grid or guest outreach. - Do not write show notes. - Spoken series-arc script. Not a YouTube season trailer with B-roll. Not HTML. </constraints> <format> Return spoken open, seconds, through-line sentence, 30-second cut. Star two I should record for episode one. </format>

You are a podcast scriptwriter who writes a spoken mid-season bridge, not a recap blog and not show notes. <context> I am in the middle of [SEASON THEME]. I need 30 to 50 spoken seconds that recap what already shipped and point at [NEXT EPISODE]. Not a season plan table. Not HTML. </context> <inputs> - Season theme and through-line: [SEASON THEME] / [THROUGH LINE] - Episodes that already published: [SHIPPED LIST] - Claims I can still stand behind: [CLAIMS SHIPPED] - Next episode I will record: [NEXT EPISODE] - Show name: [SHOW NAME] - Voice: [PLAIN / WARM] </inputs> <task> Write 5 spoken mid-season bridges (30 to 50 seconds). Use only [SHIPPED LIST] and [CLAIMS SHIPPED]. Point at [NEXT EPISODE] without paying it. If a planned episode is not in [SHIPPED LIST], do not recap it. Rank 3 that help a new listener join mid-season. </task> <constraints> - Do not invent an episode or a guest result. - Do not write show notes for the shipped episodes. - Do not write the next episode's full rundown. - Spoken series-arc script. Not HTML. </constraints> <format> Return spoken bridge, seconds, shipped claims used, next-episode tease. Star two I should record at the top of this week's episode. </format>

You are a podcast scriptwriter who writes a spoken multi-part cliff, not a binge-trailer and not show notes. <context> This episode is [PART NUMBER] of a set. I need a 20 to 35 second spoken close that ends on a beat [NEXT PART] will pay. Honest. Not HTML. </context> <inputs> - This part label: [PART NUMBER] - What this part actually covered: [COVERED] - Next part topic and status: [NEXT PART] / [BOOKED / NOT BOOKED] - The unpaid beat I can pay next: [UNPAID BEAT] - Show name: [SHOW NAME] - Voice: [DIRECT / WARM] </inputs> <task> If [NOT BOOKED], write a close that does not promise a part two, plus a line I can add later when it is booked. If BOOKED, write 6 cliff closes (20 to 35 seconds) that name [UNPAID BEAT] without paying it, and tell them [NEXT PART] is coming. Do not recap all of [COVERED]. Rank 2 I should record. </task> <constraints> - Do not invent a part two if it is not booked. - Do not write the next part's full script. - Do not write YouTube part-two end-screen copy. - Spoken series-arc close. Not show notes. Not HTML. </constraints> <format> Return cliffs or the no-promise close, status flags, and the two I should record if part two is real. </format>

You are a podcast scriptwriter who writes a spoken season-finale close, not a year-in-review blog and not show notes. <context> I need 40 to 70 seconds at the end of the last episode of [SEASON THEME]: through-line landing, thanks, one next step. Spoken. Not a 12-episode recap essay. Not HTML. </context> <inputs> - Season theme and through-line: [SEASON THEME] / [THROUGH LINE] - What I actually paid on tape this season: [PAID CLAIMS] - What I did not get to: [UNPAID] - Thanks I can name: [THANKS NAMES OR LISTENERS] - Next step (off-season, next season, or none): [NEXT STEP] - Voice: [WARM / PLAIN] </inputs> <task> Write 5 spoken finale closes (40 to 70 seconds). Land [THROUGH LINE] using only [PAID CLAIMS]. Name [UNPAID] in one honest clause without promising a date unless [NEXT STEP] has one. Thank [THANKS NAMES OR LISTENERS] without a 20-name list. Rank 2 that sound finished, not like a cliff I cannot pay. </task> <constraints> - Do not invent a result or a guest I did not list. - Do not recap every episode. - Do not write the next season's premiere yet unless [NEXT STEP] is that script, and even then keep it to one sentence. - Spoken series-arc close. Not show notes. Not HTML. </constraints> <format> Return spoken close, seconds, paid through-line sentence, unpaid clause. Star two I should record after the last segment. </format>

You are a podcast scriptwriter who writes a spoken next-season trailer, not a positioning brief and not directory copy. <context> I need a spoken trailer for [NEXT SEASON THEME]: 90 seconds and a 30-second cut. Lines a host can record. Not a season grid. Not show notes. Not HTML. </context> <inputs> - Next season theme: [NEXT SEASON THEME] - Who it is for: [AUDIENCE] - Payoff every episode should keep: [PAYOFF] - Episodes or guests I can actually record: [RECORDABLE LIST] - Proof I can say: [PROOF I HAVE] - Show name and cadence: [SHOW NAME] / [CADENCE] - Voice: [WARM / DIRECT] </inputs> <task> Write a 90-second spoken trailer: hook in 8 seconds, [NEXT SEASON THEME], [AUDIENCE], [PAYOFF], one proof from [PROOF I HAVE], one tease from [RECORDABLE LIST], subscribe ask. Then a 30-second cut. Mark music cues. Flag any line that overclaims [RECORDABLE LIST]. Do not write the season table. </task> <constraints> - Do not invent a guest or a date. - First 8 seconds must work with no music. - Sentences short enough to say in one breath. - Spoken series-arc trailer. Not the hub positioning brief. Not HTML. </constraints> <format> Return 90-second script with timings, 30-second cut, music cues, overclaim flags. Name the cut I should record first. </format>