Claude Prompt Library

30 Claude Prompts for Course Outlines You Still Teach

30 copy-paste prompts

XML prompts for outcomes, modules, lessons, exercises, a launch plan, and a sales-page outline. Claude writes the curriculum. You still record and teach.

In short: This page contains 30 copy-paste ready prompts, organized into 6 categories with a description and pro tip for each. The first 5 prompts are free instantly, no signup needed. Hand-curated and tested by the AI Academy team.

Louis Corneloup
By Louis Corneloup ยท Founder, Techpresso
Last updated ยทHand-curated & tested by the AI Academy team

Outcomes

5 prompts

Write Measurable Course Outcomes

1/30

โœจ What it does

Claude writes 4 to 8 measurable outcomes for [COURSE TOPIC] aimed at [LEARNER] inside [TIME BUDGET], each with one verb and a proof you can film. You still record the lessons and teach from those outcomes.

<context> I am writing a course outline, not recording yet and not building an HTML page. I need measurable outcomes I can teach and film. Do not invent a student result, testimonial, or skill I cannot demonstrate. </context> <inputs> - Course topic: [COURSE TOPIC] - Who will take it: [LEARNER] - Starting skill: [STARTING SKILL] - Total teaching time I can defend: [TIME BUDGET] - What I can already teach live: [SKILLS I CAN TEACH] - Notes I already have: [PASTE NOTES] </inputs> <task> Write 4 to 8 course outcomes for [COURSE TOPIC]. For each: (1) one measurable verb, (2) the artifact or performance [LEARNER] will produce, (3) how I can observe or film it, (4) which line in [PASTE NOTES] or [SKILLS I CAN TEACH] supports it, (5) a drop tag if [TIME BUDGET] cannot carry it. End with a one-sentence course promise that only uses kept outcomes. </task> <constraints> - Do not invent a student quote, completion rate, or job placement number. - Do not write an outcome I cannot teach from [SKILLS I CAN TEACH]. - Prefer fewer honest outcomes over a complete-looking list. - If a proof is missing, write ASK ME FOR PROOF. Do not fill it. - This is a written outline. I will record and teach later. </constraints> <format> Numbered outcomes with verb, artifact, proof, source tag, keep/drop. Then the course promise and an ASK ME FOR PROOF list. </format>

๐Ÿ’ก

Pro tip: If you cannot film or watch the proof, the outcome is a slogan. Cut it before you name modules.

Map Outcomes to Learner Jobs

2/30

โœจ What it does

Claude maps each outcome for [COURSE TOPIC] to a job already named in [PASTE NOTES] and leaves empty rows where no job exists. You still teach the map live and record the examples from your own work.

<context> I need each outcome tied to a job [LEARNER] already has. A job is work they already do, not a dream title. Use only jobs in my paste. This is a map, not a sales page. </context> <inputs> - Course topic: [COURSE TOPIC] - Learner: [LEARNER] - Outcome list I already have: [OUTCOME LIST] - Jobs they already do, from notes: [PASTE NOTES] - Jobs I will not claim: [OFF-LIMITS JOBS] </inputs> <task> Map each item in [OUTCOME LIST] to one job in [PASTE NOTES]: the job in their words, the Monday-morning use, and a scene I can teach. If an outcome has no job in the paste, tag NO JOB IN NOTES and do not invent one. Honor [OFF-LIMITS JOBS]. </task> <constraints> - Do not invent a job title, salary, or hiring manager. - Do not upgrade a side task into a career change. - Do not write a testimonial that maps the job. - If [PASTE NOTES] are thin, return a shorter map and a question list I must ask buyers. </constraints> <format> Table: outcome, job, Monday use, teachable scene, tag. Then NO JOB IN NOTES rows and buyer questions. </format>

๐Ÿ’ก

Pro tip: Paste jobs from sales calls, not from a persona doc. Invented jobs become modules you cannot defend on a live call.

Sequence Outcomes by Difficulty

3/30

โœจ What it does

Claude sequences [OUTCOME LIST] for [COURSE TOPIC] from first win to capstone inside [TIME BUDGET] and will not invent extra weeks. You still record the modules in that order and teach the first win yourself.

<context> I need a teaching order for outcomes I already wrote. First win early. Capstone last. Do not add outcomes. Do not invent a week count I did not give. </context> <inputs> - Course topic: [COURSE TOPIC] - Outcome list: [OUTCOME LIST] - Starting skill: [STARTING SKILL] - Time budget: [TIME BUDGET] - First win I want them to feel: [FIRST WIN] - Hardest skill I will teach: [HARDEST SKILL] </inputs> <task> Sequence [OUTCOME LIST] from first win to capstone. For each step: why it is next, what [STARTING SKILL] must already hold, what breaks if I skip it, and a time share that sums to [TIME BUDGET]. If [FIRST WIN] is later than step two, say so and propose a smaller first win from the list I already have. </task> <constraints> - Do not add a new outcome to make the ladder pretty. - Do not invent a 12-week structure if [TIME BUDGET] is shorter. - If [HARDEST SKILL] has no prerequisite on the list, flag the hole. Do not invent a prep module. - Keep the first win teachable in one sitting. </constraints> <format> Ordered ladder with time shares, skip risks, then a first-win note and any hole. </format>

๐Ÿ’ก

Pro tip: Put a finishable win in session one. A ladder that starts with theory is how students vanish before module two.

Cut Outcomes You Cannot Teach

4/30

โœจ What it does

Claude cuts outcomes from [OUTCOME LIST] that [TIME BUDGET] and [SKILLS I CAN TEACH] cannot support, and keeps a short list you can defend. You still record only what you can teach and drop the rest before launch.

<context> My outline is bloated. I need an honesty cut: what I can teach and film in the time I have. Do not replace a cut outcome with a trendier one I did not name. </context> <inputs> - Course topic: [COURSE TOPIC] - Outcome list: [OUTCOME LIST] - Time budget: [TIME BUDGET] - Skills I can teach live: [SKILLS I CAN TEACH] - Assets I already have: [ASSET LIST] - Tempting extras: [TEMPTING TOPICS] </inputs> <task> Sort [OUTCOME LIST] into KEEP, CUT, and PARK. KEEP only if I can teach it from [SKILLS I CAN TEACH] and [ASSET LIST] inside [TIME BUDGET]. CUT if I cannot demonstrate it. PARK tempting extras for a later course. Write a shorter promise sentence from KEEP only. </task> <constraints> - Do not invent a guest teacher, tool license, or dataset to save a CUT row. - Do not keep an outcome because it looks good on a sales page. - Time shares for KEEP must sum to [TIME BUDGET] or less. - If KEEP is under three outcomes, say the course is not ready and stop. </constraints> <format> KEEP / CUT / PARK tables, then the shorter promise, then a not-ready note if KEEP is too thin. </format>

๐Ÿ’ก

Pro tip: Share the CUT table with sales before they write the page. Those rows are the promises they will try to restore.

Outcome-to-Proof Checklist

5/30

โœจ What it does

Claude turns [OUTCOME LIST] into a proof checklist for [COURSE TOPIC] with a demo, artifact, or quiz you can actually run. You still record each proof and teach the standard you will grade.

<context> Each outcome needs a proof I can run on camera or in a live room. This is a QA list, not a certificate design and not a sales stack. Empty proof stays empty. </context> <inputs> - Course topic: [COURSE TOPIC] - Outcome list: [OUTCOME LIST] - Proof I already have: [PASTE PROOF] - Format I can run: [QUIZ / DEMO / ARTIFACT / LIVE CRIT] - Time I can spend on proof per outcome: [PROOF TIME] </inputs> <task> For each outcome, write one proof: the task, the passing look, the fail look, the format from [FORMAT], and a source tag in [PASTE PROOF]. If no proof exists, write ASK ME FOR PROOF. Do not invent a student sample or a score distribution. </task> <constraints> - Do not invent a before/after metric. - Do not write a certificate line or a badge name. - Stay inside [PROOF TIME]. - Prefer one harsh honest proof over three decorative quizzes. </constraints> <format> One card per outcome: task, pass, fail, format, source or ASK ME FOR PROOF. Then a minimum proof set I should record first. </format>

๐Ÿ’ก

Pro tip: Film one passing sample yourself before launch. If you cannot produce the artifact, students will not either.

Modules

5 prompts

Group Outcomes into Modules

6/30

โœจ What it does

Claude groups [OUTCOME LIST] into modules for [COURSE TOPIC] with one job per module and a time share of [TIME BUDGET]. You still record each module and teach the live or video sessions.

<context> I have outcomes. I need modules with one job each. This is a written curriculum map, not a designed course homepage and not a lesson script. </context> <inputs> - Course topic: [COURSE TOPIC] - Outcome list: [OUTCOME LIST] - Learner: [LEARNER] - Time budget: [TIME BUDGET] - Max modules I can record: [MAX MODULES] - Notes: [PASTE NOTES] </inputs> <task> Group [OUTCOME LIST] into at most [MAX MODULES] modules. For each: one-sentence job, outcomes it owns, a time share that sums to [TIME BUDGET], the first lesson title, and the exit proof. If an outcome fits nowhere, leave it UNPLACED. Do not invent a bonus module. </task> <constraints> - One job per module. Topic labels are not jobs. - Do not add outcomes I did not list. - Do not create a welcome module unless an outcome requires orientation. - If [MAX MODULES] is too small, say which outcomes must drop. Do not squash five jobs into one module. </constraints> <format> Module cards (job, outcomes, time, first lesson, exit proof), then UNPLACED outcomes, then a one-line course spine. </format>

๐Ÿ’ก

Pro tip: If a module needs two jobs, split it. Double-job modules are where recordings ramble and students skip.

Name Modules as Promises

7/30

โœจ What it does

Claude renames modules for [COURSE TOPIC] as promises [LEARNER] can finish, not as topic labels from [MODULE LIST]. You still teach under those names and record titles you will actually use.

<context> My module names read like a syllabus. I need promise names [LEARNER] can finish, not topic labels. Do not invent a punchy name that claims a result I cannot teach. </context> <inputs> - Course topic: [COURSE TOPIC] - Learner: [LEARNER] - Module list with jobs: [MODULE LIST] - Words I never use: [BANNED PHRASES] - Platform character limit if I have one: [NAME LIMIT OR NONE] </inputs> <task> Rename each item in [MODULE LIST]. Give 2 name options per module: a short promise and a longer classroom title. Each name must match the module job and stay inside what I can teach. Flag any option that overclaims. Honor [BANNED PHRASES] and [NAME LIMIT]. </task> <constraints> - Do not use unlock, masterclass, or secret unless I already use those words. - Do not name a result I cannot prove in that module. - Do not number modules as Module 1 unless I asked. - Keep names speakable on camera in one breath. </constraints> <format> Table: old name, short promise, classroom title, overclaim flag. Then a recommended set in teaching order. </format>

๐Ÿ’ก

Pro tip: Say the name out loud. If it sounds like a textbook chapter, rewrite it as the job the student leaves able to do.

Sequence Modules and Dependencies

8/30

โœจ What it does

Claude sequences [MODULE LIST] for [COURSE TOPIC] so each module has a prerequisite you already listed, and flags skips that break the course. You still record in that order and teach the unlock live if a student jumps ahead.

<context> I need a module order with real dependencies. If a student skips a module, I need to know what breaks. Do not invent a prerequisite I did not list. </context> <inputs> - Course topic: [COURSE TOPIC] - Module list: [MODULE LIST] - Outcome list: [OUTCOME LIST] - Starting skill: [STARTING SKILL] - Optional modules I might skip: [OPTIONAL MODULES OR NONE] </inputs> <task> Order [MODULE LIST]. For each module: required prior modules, the skill that must already be true, what breaks if skipped, and whether it can be optional. If two modules have no dependency, say they can swap. Do not add a foundation module I did not list. </task> <constraints> - Do not invent a tool-setup module to paper over a thin start. - If [STARTING SKILL] cannot enter module one, flag the hole. Do not invent a prep course. - Keep optional tags rare. Most skippable modules should stay PARK, not fake-optional. </constraints> <format> Ordered list with prior, must-already-be-true, skip risk, optional tag. Then swap pairs and a hole list. </format>

๐Ÿ’ก

Pro tip: If module three needs a file from module one, say so in the outline. Hidden file dependencies are how cohort week three dies.

Module Time and Scope Budget

9/30

โœจ What it does

Claude budgets minutes for each module in [MODULE LIST] so they sum to [TIME BUDGET] without inventing extra weeks or studio days. You still record to that clock and teach the cuts when a module runs long.

<context> I will record and teach this myself. Minutes must add up. Do not invent extra weeks, live calls, or a TA I do not have. </context> <inputs> - Course topic: [COURSE TOPIC] - Module list: [MODULE LIST] - Time budget: [TIME BUDGET] - Lesson length I can record: [LESSON LENGTH] - Live calls I can host: [LIVE CALLS OR NONE] - Days I have to record: [DAYS I HAVE] </inputs> <task> Budget each module: teaching minutes, number of lessons at [LESSON LENGTH], practice minutes, and whether a live call is required. Sums must fit [TIME BUDGET] and [DAYS I HAVE]. If the list does not fit, cut scope, do not add days. Note modules that need a live call I did not offer. </task> <constraints> - Do not invent a 12-week calendar if [TIME BUDGET] is six hours. - Do not add office hours I did not list. - Recording time is not the same as student time. Show both if they differ. - If a module needs more than I can film in [DAYS I HAVE], mark CUT OR SPLIT. </constraints> <format> Budget table (teach, lessons, practice, live, record days), totals, then CUT OR SPLIT rows. </format>

๐Ÿ’ก

Pro tip: Budget recording days, not just student hours. A six-hour course that needs twenty days in the studio will miss the launch date.

What This Course Will Not Cover

10/30

โœจ What it does

Claude writes a not-in-this-course list for [COURSE TOPIC] from [TEMPTING TOPICS] and [TIME BUDGET] so the outline stays finishable. You still teach the boundary in lesson one and record the scope line on the sales page.

<context> I need a hard not-in-this-course list so the outline and the sales page stay honest. Do not fill the gaps with a bonus module I will not record. </context> <inputs> - Course topic: [COURSE TOPIC] - Kept outcomes: [OUTCOME LIST] - Tempting extras: [TEMPTING TOPICS] - Time budget: [TIME BUDGET] - Adjacent courses I might sell later: [LATER COURSES OR NONE] - Audience confusion I already hear: [CONFUSION I HEAR] </inputs> <task> Write a scope memo: (1) in-scope in one sentence, (2) adjacent topics I will mention in one slide only, (3) topics I will not teach and why, (4) a spoken scope line for lesson one, (5) a sales-page line that matches. Park [TEMPTING TOPICS] that belong in [LATER COURSES]. Answer [CONFUSION I HEAR] without adding scope. </task> <constraints> - Do not invent a later course title I did not list. - Do not keep a tempting topic because a competitor teaches it. - The spoken line and the sales line must match. - Specific excludes, not everything else. </constraints> <format> Five headed sections, then a one-line scope I can put above the outline. </format>

๐Ÿ’ก

Pro tip: Paste the last sales-call objection. Claude is better at cutting when it can see the extra topic buyers keep asking for.

Lessons

5 prompts

Break a Module into Lessons

11/30

โœจ What it does

Claude breaks [MODULE NAME] into lessons for [LEARNER] that fit [LESSON LENGTH] and serve [MODULE OUTCOME]. You still record each lesson and teach the transitions in your own words.

<context> I need one module turned into lessons I can record. Each lesson has one job. This is a written breakdown, not a full script and not HTML. </context> <inputs> - Course topic: [COURSE TOPIC] - Module name and job: [MODULE NAME] - Module outcome: [MODULE OUTCOME] - Learner: [LEARNER] - Lesson length I can record: [LESSON LENGTH] - Notes: [PASTE NOTES] </inputs> <task> Break [MODULE NAME] into lessons that fit [LESSON LENGTH]. For each: one job, open in one line, teach block, practice, close, and the file name I should use. Lessons must sum to the module job. If [PASTE NOTES] cannot support a lesson, write NEEDS NOTES instead of filling it. </task> <constraints> - Do not write the spoken script in this pass. - Do not add a recap lesson that restates the module. - Do not invent an example that is not in [PASTE NOTES]. - Prefer fewer lessons that finish over a pile of 8-minute fragments. </constraints> <format> Numbered lessons with job, open, teach, practice, close, file name. Then NEEDS NOTES rows. </format>

๐Ÿ’ก

Pro tip: One job per file. If a lesson needs two jobs, you will either rush the teach block or skip practice.

Write One Lesson Outline

12/30

โœจ What it does

Claude outlines one lesson on [LESSON TOPIC] with open, teach, practice, and close that fit [LESSON LENGTH] and use [EXAMPLE I HAVE]. You still record that lesson and teach it from the outline, not from a dumped script.

<context> I need a single lesson outline I can teach from. Not a teleprompter dump. Not a designed slide deck. I will record it later in my voice. </context> <inputs> - Lesson topic: [LESSON TOPIC] - Module outcome it serves: [MODULE OUTCOME] - Lesson length: [LESSON LENGTH] - Learner: [LEARNER] - Example I can actually show: [EXAMPLE I HAVE] - Notes: [PASTE NOTES] </inputs> <task> Outline one lesson: (1) open that names the job, (2) teach beats with talking points from [PASTE NOTES], (3) the worked [EXAMPLE I HAVE], (4) a practice the student can finish before the close, (5) a close that restates [MODULE OUTCOME], (6) a minute map that sums to [LESSON LENGTH]. Leave a hole where notes are thin. </task> <constraints> - Do not invent a case, screenshot, or number I did not paste. - Do not write every spoken sentence. Talking points only. - If [EXAMPLE I HAVE] cannot carry the teach block, say so and stop. - No welcome-back or like-and-subscribe filler. </constraints> <format> Minute map, then open / teach / example / practice / close, then holes I must fill before I record. </format>

๐Ÿ’ก

Pro tip: If the example is one you have not run, stop. A lesson outline with a fake walkthrough will collapse on camera.

Opening Hook and Lesson Close

13/30

โœจ What it does

Claude writes a spoken open and close for [LESSON TOPIC] that pay [MODULE OUTCOME] and use only [PROOF I CAN SHOW]. You still record the open on camera and teach the close live.

<context> I need a spoken open and close for one lesson. I will say these lines on camera or live. Do not invent a student story to make the hook land. </context> <inputs> - Lesson topic: [LESSON TOPIC] - Module outcome: [MODULE OUTCOME] - Learner: [LEARNER] - Proof I can show in the first 20 seconds: [PROOF I CAN SHOW] - Voice: [DIRECT / WARM / CALM] - Lesson length: [LESSON LENGTH] </inputs> <task> Write 5 open options (8 to 20 seconds) and 3 close options (15 to 30 seconds). Each open must name the job and use only [PROOF I CAN SHOW]. Each close must restate [MODULE OUTCOME] and point to the practice, not to a new topic. Mark any line I cannot say in one breath. </task> <constraints> - Do not invent a student name, quote, or result. - No hey-everyone or welcome-back. - The close must not introduce a new outcome. - If [PROOF I CAN SHOW] is NONE, open on the job, not on a fake result. </constraints> <format> Ranked opens with seconds and proof tag, ranked closes, then two pairs I should record first. </format>

๐Ÿ’ก

Pro tip: If the only honest open is the job in one sentence, use it. A fake student win in the first 10 seconds will follow you into the comments.

Lesson Talking Points from Notes

14/30

โœจ What it does

Claude turns [PASTE NOTES] into talking points for [LESSON TOPIC] and leaves a hole where the notes are thin. You still record the lesson from those points and teach the examples you actually ran.

<context> I have notes I already trust. I need talking points for one lesson, not new research and not a full script. Thin notes stay thin. </context> <inputs> - Lesson topic: [LESSON TOPIC] - Notes or transcript I already have: [PASTE NOTES] - Module outcome: [MODULE OUTCOME] - Example I can show: [EXAMPLE I HAVE] - Claims I will not make: [OFF-LIMITS CLAIMS] </inputs> <task> Turn [PASTE NOTES] into numbered talking points in teaching order. Each point: the claim, the source line, the example if present, and a check I can ask out loud. If a point needs a fact I did not paste, write NEEDS NOTES. Honor [OFF-LIMITS CLAIMS]. </task> <constraints> - Do not invent a study, tool tip, or statistic. - Do not expand a one-line note into a three-act story. - Do not write the full spoken paragraph unless a line is a definition I must say exactly. - If the notes cannot support [MODULE OUTCOME], say the lesson is not ready. </constraints> <format> Numbered talking points with source tags, then NEEDS NOTES, then a not-ready note if the outcome cannot stand. </format>

๐Ÿ’ก

Pro tip: Paste the notes you would teach from tomorrow. Memory notes are how a confident false step gets into the recording.

Lesson Runtime and Recording Plan

15/30

โœจ What it does

Claude builds a recording plan for [LESSON LIST] with runtime, B-roll you can film, and a retake rule that fits [DAYS I HAVE]. You still record each take and teach the live version if a file needs a fix.

<context> I will record these lessons myself. I need a shoot plan: order, runtime, retakes, and B-roll I can actually film. Do not invent a studio, editor, or stock library. </context> <inputs> - Lesson list: [LESSON LIST] - Lesson length I want on the student side: [LESSON LENGTH] - Days I have: [DAYS I HAVE] - B-roll I can film: [B-ROLL I CAN FILM] - Tools on my desk: [TOOLS I HAVE] - Face on camera: [YES / NO] </inputs> <task> Build a recording plan: shoot order, target runtime, buffer for retakes, B-roll from [B-ROLL I CAN FILM] only, and a same-day checklist. If a lesson needs B-roll I do not have, mark TYPE ONLY. Sums must fit [DAYS I HAVE]. Do not add a shoot day. </task> <constraints> - Do not invent a second camera, a set, or an editor. - Do not request stock footage. - If [FACE ON CAMERA] is NO, plan voice-over plus the B-roll I named. - Prefer fewer honest files over a shot list I cannot finish. </constraints> <format> Day-by-day shoot table, TYPE ONLY lessons, then a same-day checklist I can print. </format>

๐Ÿ’ก

Pro tip: Schedule retake time on the same day. A lesson you swear you will fix next week is the one that ships with the stumble still in it.

These prompts give you the what. Tutorials give you the why.

Learn when to use extended thinking, how to build Claude Projects, and workflows that compound. 300+ tutorials and growing.

Try AI Academy Free

Exercises

5 prompts

Design a Practice Exercise

16/30

โœจ What it does

Claude designs one practice exercise for [LESSON TOPIC] that [LEARNER] can finish in [EXERCISE TIME] with assets you already have. You still run the exercise in class and record the walkthrough yourself.

<context> I need one practice exercise students can finish after a lesson. I will run it live or on video. Do not invent a dataset, tool login, or client file I do not have. </context> <inputs> - Lesson topic: [LESSON TOPIC] - Outcome it proves: [OUTCOME] - Learner: [LEARNER] - Time they have: [EXERCISE TIME] - Assets I can give them: [ASSET LIST] - Tools they already have: [TOOLS THEY HAVE] </inputs> <task> Design one exercise: goal in one sentence, steps, the file or prompt they start with (from [ASSET LIST] only), the done look, a common fail, and a 2-minute teacher walkthrough outline. If an asset is missing, write NEEDS ASSET and stop. Stay inside [EXERCISE TIME]. </task> <constraints> - Do not invent a spreadsheet, API key, or sample client. - Do not require a paid tool that is not in [TOOLS THEY HAVE]. - Do not write a group exercise if I did not say I host a cohort. - The walkthrough must be something I can record once. </constraints> <format> Goal, steps, start file, done look, fail, walkthrough outline, then NEEDS ASSET if blocked. </format>

๐Ÿ’ก

Pro tip: Do the exercise once with the same files students get. If you needed a hidden tab, the exercise is not ready.

Build a Worksheet from a Lesson

17/30

โœจ What it does

Claude writes a worksheet for [LESSON TOPIC] from [PASTE NOTES] with blanks, a worked [EXAMPLE I HAVE], and no invented data. You still teach the worked example and record the fill-in as students work.

<context> I need a worksheet that follows one lesson. Blanks, a worked example I already have, and no invented numbers. This is copy I can paste into a doc, not a designed PDF. </context> <inputs> - Lesson topic: [LESSON TOPIC] - Notes: [PASTE NOTES] - Worked example I can show: [EXAMPLE I HAVE] - Outcome: [OUTCOME] - Time: [EXERCISE TIME] - Blanks I want: [BLANK COUNT] </inputs> <task> Write a worksheet: title, one-line why, the worked [EXAMPLE I HAVE], then [BLANK COUNT] student blanks that mirror the example, and a self-check of 4 questions. Use only facts in [PASTE NOTES]. If the example cannot support a blank, write NEEDS EXAMPLE DATA. </task> <constraints> - Do not invent a data row to make the sheet look complete. - Do not add a second example I did not paste. - Keep instructions short enough to read on a phone. - This is a written worksheet, not HTML. </constraints> <format> Worksheet copy in teaching order, then NEEDS EXAMPLE DATA rows, then a 30-second teacher note. </format>

๐Ÿ’ก

Pro tip: If a blank needs a number you do not have, leave it empty. A tidy invented row teaches the wrong math.

Write a Capstone Project Brief

18/30

โœจ What it does

Claude writes a capstone brief for [COURSE TOPIC] that proves [OUTCOME LIST] using only deliverables [LEARNER] can make in [TIME BUDGET]. You still teach the brief live and record a walkthrough of a real sample you made.

<context> I need a capstone brief that proves the kept outcomes. Students must finish it with time and tools I already named. Do not invent a client, a dataset, or a portfolio piece I will not review. </context> <inputs> - Course topic: [COURSE TOPIC] - Outcome list I kept: [OUTCOME LIST] - Learner: [LEARNER] - Time they have after the lessons: [TIME BUDGET] - Tools they have: [TOOLS THEY HAVE] - Sample I can show: [SAMPLE I MADE OR NONE] </inputs> <task> Write a capstone brief: promise in one sentence, deliverables mapped to [OUTCOME LIST], steps, timebox, tools, a done look, and what I will not review. If an outcome cannot appear in the deliverable, tag UNPROVABLE and recommend dropping it from the course promise. If [SAMPLE I MADE] is NONE, list the sample I must record first. </task> <constraints> - Do not invent a partner company or a live client. - Do not require a tool I did not list. - Do not write a 40-hour capstone if [TIME BUDGET] is four hours. - The brief must be teachable in one live session. </constraints> <format> Brief, outcome-to-deliverable map, UNPROVABLE rows, then the sample I must record. </format>

๐Ÿ’ก

Pro tip: Ship the brief only after you have made the sample. A capstone you have not done is a support queue, not a course.

Rubric for an Exercise

19/30

โœจ What it does

Claude writes a rubric for [EXERCISE NAME] with levels you can score in one pass, tied to [OUTCOME] and [PASTE NOTES]. You still grade with that rubric and teach the standard in a recorded review.

<context> I need a rubric I can score in one pass, live or on a recording. Levels must be visible in the student file. Do not invent a grade curve or a student sample score. </context> <inputs> - Exercise name: [EXERCISE NAME] - Outcome: [OUTCOME] - Notes or brief: [PASTE NOTES] - Levels I use: [LEVELS] - Time I have to score one file: [SCORE TIME] - Must-fail errors: [MUST FAIL] </inputs> <task> Write a rubric with [LEVELS] (for example miss / approaching / pass / strong). Each cell: what I will see in the file, in one line. Add a must-fail list from [MUST FAIL]. Keep the whole rubric scorable inside [SCORE TIME]. No points unless I asked for points. </task> <constraints> - Do not invent a student name or an example score. - Do not add a criterion that is not in [PASTE NOTES] or [OUTCOME]. - Avoid taste words (creative, professional) unless you define a visible tell. - I will teach this rubric on camera. Keep the language I can say out loud. </constraints> <format> Rubric table, must-fail list, then a 60-second spoken version I can record. </format>

๐Ÿ’ก

Pro tip: Read each cell and ask whether you would see it in a file in 90 seconds. If not, the cell is a vibe, not a rubric.

Reusable Feedback Comments

20/30

โœจ What it does

Claude drafts reusable feedback lines for [EXERCISE NAME] at each level in [RUBRIC] without inventing a student name or score. You still record the feedback video and teach the fix in office hours.

<context> I need reusable feedback lines for one exercise, by rubric level. I will paste them and then record a short fix. Do not invent a student, a score, or a private detail. </context> <inputs> - Exercise name: [EXERCISE NAME] - Rubric I already have: [RUBRIC] - Outcome: [OUTCOME] - Common fails I already see: [COMMON FAILS] - Voice: [DIRECT / WARM] - Words I never use: [BANNED PHRASES] </inputs> <task> For each level in [RUBRIC], write 3 comments I can paste: what is true, what to fix next, and a pointer back to the lesson. Add 3 comments for each item in [COMMON FAILS]. Leave a [NAME] token I will replace. Do not write a full video script. </task> <constraints> - Do not invent a student story. - Do not praise work I have not seen. - Honor [BANNED PHRASES]. - Keep each comment under 60 words so I can say it on a Loom. </constraints> <format> Comments grouped by level and by common fail, each with a one-line record cue. </format>

๐Ÿ’ก

Pro tip: Replace [NAME] and one concrete line from their file before you send. A pasted comment with no file detail feels like a bot.

Launch

5 prompts

Pre-Launch Asset Checklist

21/30

โœจ What it does

Claude builds a pre-launch checklist for [COURSE TOPIC] on [PLATFORM] from [MODULE LIST] with only assets you said you have. You still record the missing lessons and teach week one before you open the cart.

<context> I am close to opening the cart. I need a checklist of assets I said I have, not a motivational launch plan. Missing files stay missing. </context> <inputs> - Course topic: [COURSE TOPIC] - Platform: [PLATFORM] - Module list: [MODULE LIST] - Assets I already have: [ASSET LIST] - Offer dates: [OFFER DATES] - Live calls I will host: [LIVE CALLS OR NONE] </inputs> <task> Build a pre-launch checklist: lessons recorded, exercises attached, capstone brief, welcome, sales page outline, emails, and live-call slots. Mark each item HAVE, MISSING, or NOT NEEDED. If [OFFER DATES] arrive before MISSING lesson files, say DELAY. Do not invent an asset to look ready. </task> <constraints> - Do not add a community, a Slack, or a bonus I did not list. - Do not mark a module HAVE unless it is in [ASSET LIST]. - Platform setup steps must match [PLATFORM] only. - If live calls are NONE, do not schedule a kickoff I will not host. </constraints> <format> Checkbox groups, then MISSING blockers, then a GO / DELAY line tied to [OFFER DATES]. </format>

๐Ÿ’ก

Pro tip: A DELAY on recorded lessons is a gift. Opening the cart on a half-built outline is how refunds start in week one.

Waitlist and First-Week Plan

22/30

โœจ What it does

Claude writes a waitlist and first-week plan for [COURSE TOPIC] aimed at [LEARNER] with emails you can send and a live session you can host. You still send those emails and teach the first live call yourself.

<context> I need a waitlist plan and the first week after someone buys. Emails and one live session I can actually host. Do not invent a 12-email sequence I will not send. </context> <inputs> - Course topic: [COURSE TOPIC] - Learner: [LEARNER] - Platform: [PLATFORM] - First week goal: [FIRST WEEK GOAL] - Emails I will send: [EMAIL COUNT] - Live session I can host: [LIVE CALLS OR NONE] - Dates: [OFFER DATES] </inputs> <task> Write: (1) waitlist emails up to [EMAIL COUNT], (2) the day-zero access email, (3) a first-week calendar that hits [FIRST WEEK GOAL], (4) one live session outline if I named a call, (5) a do-not-send list. Every email is an outline with subject, 5 bullets, and a CTA I can keep. No invented proof. </task> <constraints> - Do not invent a testimonial or a seat count for urgency. - Do not write 12 emails if I said 4. - If live calls are NONE, the first week must work asynchronously. - Subjects must be sendable. No fake Re: or Fwd:. </constraints> <format> Waitlist outlines, day-zero email, first-week calendar, live outline or NONE, then do-not-send. </format>

๐Ÿ’ก

Pro tip: Write the day-zero email before the waitlist teases. If access is messy, the teases only raise refund risk.

Recording Calendar from the Outline

23/30

โœจ What it does

Claude turns [MODULE LIST] into a recording calendar that fits [DAYS I HAVE] and [LESSON LENGTH] without inventing extra studio days. You still record on those days and teach the live makeup if a file fails.

<context> I have a module and lesson list. I need a calendar I can shoot against. Do not invent extra studio days, an editor, or a second instructor. </context> <inputs> - Module list: [MODULE LIST] - Lesson list: [LESSON LIST] - Lesson length: [LESSON LENGTH] - Days I have: [DAYS I HAVE] - Hours I can record per day: [HOURS PER DAY] - Hard dates I cannot move: [HARD DATES] </inputs> <task> Turn the lists into a recording calendar that fits [DAYS I HAVE] and [HOURS PER DAY]. Batch similar lessons. Put retake blocks on the same day. Honor [HARD DATES]. If the math does not fit, cut lessons or shorten [LESSON LENGTH]. Do not add a day. </task> <constraints> - Do not assume I can record 8 finished hours in one day. - Do not schedule a lesson I did not list. - Show student runtime and estimated shoot time as two numbers. - If I must cut, name the lesson and the outcome it orphans. </constraints> <format> Calendar by day, shoot vs student time, then cuts required to fit. </format>

๐Ÿ’ก

Pro tip: Count shoot hours as twice the student runtime until you have a real average. Optimistic calendars miss the launch.

Student Onboarding Outline

24/30

โœจ What it does

Claude outlines onboarding for [COURSE TOPIC] on [PLATFORM] with a start lesson, a tool check, and a first win in [FIRST WEEK GOAL]. You still record the welcome and teach the first win on the call.

<context> I need an onboarding outline so a new student can reach the first win without a support thread. Written steps for [PLATFORM]. I will record the welcome myself. </context> <inputs> - Course topic: [COURSE TOPIC] - Platform: [PLATFORM] - First week goal: [FIRST WEEK GOAL] - Tools they must install: [TOOLS THEY HAVE] - Welcome I can record: [WELCOME LENGTH] - Support I actually offer: [SUPPORT I OFFER] </inputs> <task> Outline onboarding: welcome video beats that fit [WELCOME LENGTH], tool check, where to click on [PLATFORM], the first lesson, the first exercise, and how [SUPPORT I OFFER] works. Include a stuck path with 3 replies I can reuse. Do not promise a Slack, a community, or office hours I did not list. </task> <constraints> - Do not invent a setup I cannot screenshot. - Do not add a personality quiz or a long survey. - The first win must be [FIRST WEEK GOAL], not a tour of the whole curriculum. - Keep steps numbered and short enough to film as on-screen cards. </constraints> <format> Numbered onboarding, welcome beats, stuck-path replies, then a do-not-promise list. </format>

๐Ÿ’ก

Pro tip: Onboard to the first win, not to the platform tour. Students who only learn where buttons are still have not started the course.

Launch Week Comms Plan

25/30

โœจ What it does

Claude writes a launch-week comms plan for [COURSE TOPIC] with posts, emails, and a live slot that match [OFFER DATES] and [CHANNEL LIST]. You still record the launch video and teach the AMA yourself.

<context> I need a launch-week comms outline I can actually send and record. Posts, emails, and one live slot. Do not invent a partner blast, a waitlist size, or a fake countdown. </context> <inputs> - Course topic: [COURSE TOPIC] - Offer dates: [OFFER DATES] - Channels I will use: [CHANNEL LIST] - Emails I will send: [EMAIL COUNT] - Live slot I can host: [LIVE CALLS OR NONE] - Proof I can show: [PASTE PROOF] </inputs> <task> Write a day-by-day launch week: which channel, the job of the message, 5 bullets, and the CTA. Use only [PASTE PROOF]. If a day wants a proof I do not have, mark ASK ME FOR PROOF and give a weaker honest angle. Include one live outline if I named a call. </task> <constraints> - Do not invent a seat cap, a testimonial, or a closing-in-3-hours line unless I pasted that rule. - Do not add a channel that is not in [CHANNEL LIST]. - Do not write more emails than [EMAIL COUNT]. - I will record one launch video. Outline it once, do not write five scripts. </constraints> <format> Day-by-day table, launch-video outline, live outline or NONE, then ASK ME FOR PROOF days. </format>

๐Ÿ’ก

Pro tip: If you have no proof clip, lead with the first-win lesson, not a fake student story. Buyers check.

Most people use 10% of Claude. Tutorials unlock the rest.

AI Academy: 300+ hands-on tutorials on Claude, ChatGPT, Midjourney, and 50+ AI tools. New tutorials added every week.

Start Your Free Trial

Sales Outline

5 prompts

Sales Page Outline from Outcomes

26/30

โœจ What it does

Claude outlines a sales page for [COURSE TOPIC] from [OUTCOME LIST] and [PRICE], with no invented testimonial or enrollment count. You still record the sales video and teach a live walkthrough of the curriculum.

<context> I need a sales-page outline from outcomes I can teach, not a designed landing page and not HTML. Empty proof stays empty. I will record a walkthrough later. </context> <inputs> - Course topic: [COURSE TOPIC] - Outcome list I kept: [OUTCOME LIST] - Learner: [LEARNER] - Price: [PRICE] - Proof I have: [PASTE PROOF] - Refund rule: [REFUND RULE] - Words I never use: [BANNED PHRASES] </inputs> <task> Outline the sales page in order: promise, who it is for, who it is not for, outcome list, module teaser, proof from [PASTE PROOF] only, offer at [PRICE], [REFUND RULE], FAQ stubs. For any section that wants a testimonial or enrollment count I did not paste, write ASK ME FOR PROOF. </task> <constraints> - Do not invent a student name, quote, or result. - Do not invent a seat count or a bonus. - Honor [BANNED PHRASES]. - This is an outline I will type into my page builder. Not HTML. </constraints> <format> Numbered page sections with bullets, then ASK ME FOR PROOF, then a one-line promise I can also say on camera. </format>

๐Ÿ’ก

Pro tip: Paste only proof you can show on a call. Outline holes are cheaper than a published quote you cannot defend.

Promise and Proof Stack

27/30

โœจ What it does

Claude writes a promise and proof stack for [COURSE TOPIC] using only proof in [PASTE PROOF] and marks holes as ASK ME FOR PROOF. You still record the proof clips and teach from results you can show.

<context> I need a promise line and a proof stack for the sales page and the sales video. Every proof must already sit in my paste. Do not write a bigger win to fill a third slot. </context> <inputs> - Course topic: [COURSE TOPIC] - Outcome list: [OUTCOME LIST] - Proof I have: [PASTE PROOF] - Learner: [LEARNER] - Claims I will not make: [OFF-LIMITS CLAIMS] </inputs> <task> Write 5 promise options that stay inside [OUTCOME LIST], then a proof stack of up to 3 lines from [PASTE PROOF]. Rank promises by honesty, not punch. For each proof line: the source tag and a risk note if it implies a job I did not teach. If I have fewer than 3 proofs, leave slots empty. </task> <constraints> - Do not invent a metric, logo, or student. - Do not turn a personal result into they will get this. - Honor [OFF-LIMITS CLAIMS]. - If no proof is honest enough to lead, say so and stop at a descriptive title. </constraints> <format> Ranked promises with risk notes, recommended pick, then the proof stack with empty slots left blank. </format>

๐Ÿ’ก

Pro tip: Two honest proofs beat a third invented lift. The empty slot is the clip you film next, not a sentence you guess.

Curriculum Teaser without Dumping

28/30

โœจ What it does

Claude writes a curriculum teaser for [MODULE LIST] that names promises, not a full dump of [LESSON LIST]. You still teach the teaser on a webinar and record the short version for the page.

<context> I need a curriculum teaser for the sales page and a webinar. Name promises, not every lesson file. A full dump trains people to steal the outline and skip the buy. </context> <inputs> - Course topic: [COURSE TOPIC] - Module list: [MODULE LIST] - Lesson list I must not dump: [LESSON LIST] - Learner: [LEARNER] - Minutes I have to say it: [TEASER MINUTES] </inputs> <task> Write a teaser: one line per module as a promise, a spoken version that fits [TEASER MINUTES], and a page version of 80 to 140 words. Do not list items from [LESSON LIST] unless a lesson is the first win I must name. Add a not-in-this-teaser line so I do not overshare on a live call. </task> <constraints> - Do not invent a bonus module. - Do not turn the teaser into a full syllabus. - Do not add a result the module cannot prove. - Keep the spoken version sayable without slides if needed. </constraints> <format> Module promise list, spoken teaser, page teaser, then not-in-this-teaser. </format>

๐Ÿ’ก

Pro tip: If you read the teaser and a competitor could rebuild the course, you dumped. Keep the job, hide the file list.

Objection-Ready FAQ Outline

29/30

โœจ What it does

Claude outlines sales FAQ answers for [COURSE TOPIC] from [OBJECTIONS] using only facts in [PASTE NOTES] and [REFUND RULE]. You still record FAQ answers on camera and teach the live Q&A.

<context> I need FAQ copy for the sales page and a live Q&A. Answer only with facts I pasted. Do not invent a refund story, a job outcome, or a time-to-result I did not measure. </context> <inputs> - Course topic: [COURSE TOPIC] - Objections I already hear: [OBJECTIONS] - Notes and facts I can use: [PASTE NOTES] - Refund rule: [REFUND RULE] - Time budget of the course: [TIME BUDGET] - Who it is not for: [NOT FOR] </inputs> <task> For each item in [OBJECTIONS], write: a short page answer, a longer spoken answer, and CANNOT ANSWER FROM NOTES if the paste is empty. Include [REFUND RULE], [TIME BUDGET], and [NOT FOR] only where they belong. Do not invent a fourth objection to look complete. </task> <constraints> - Do not invent a student who had the same objection. - Do not promise a job, a raise, or a client. - Do not soften [NOT FOR] to win a maybe-buyer. - Keep page answers under 70 words. </constraints> <format> FAQ pairs (page / spoken), then CANNOT ANSWER rows I must fill before publish. </format>

๐Ÿ’ก

Pro tip: The CANNOT ANSWER rows are homework, not lines to improvise on a webinar. Guessed refund stories become policy.

Offer and Guarantee Lines

30/30

โœจ What it does

Claude drafts offer and guarantee lines for [COURSE TOPIC] at [PRICE] that match [REFUND RULE] and do not invent a bonus you did not list. You still put the offer on the page and teach the terms on the sales call.

<context> I need offer and guarantee lines that match the price and refund rule I already have. Do not invent a bonus, a payment plan, or a stronger guarantee I will not honor. </context> <inputs> - Course topic: [COURSE TOPIC] - Price: [PRICE] - What is in the offer: [OFFER CONTENTS] - Refund rule: [REFUND RULE] - Bonuses I will actually deliver: [BONUS LIST OR NONE] - Words I never use: [BANNED PHRASES] </inputs> <task> Draft 4 offer lines and 3 guarantee lines. Each must name only items in [OFFER CONTENTS] and [BONUS LIST], at [PRICE], under [REFUND RULE]. Flag any line that implies a result the outline does not teach. Give one recommended pair I can put on the page and say on a sales call. </task> <constraints> - Do not invent a bonus to sweeten the line. - Do not write a lifetime or risk-free line unless [REFUND RULE] says that. - Do not add a payment plan I did not name. - Honor [BANNED PHRASES]. </constraints> <format> Offer lines with risk notes, guarantee lines with risk notes, then the recommended pair. </format>

๐Ÿ’ก

Pro tip: Read the guarantee to a friend who refunds things. If they hear a wider promise than [REFUND RULE], rewrite it before checkout goes live.

Free tool

Prompt Optimizer

Turn a rough idea into a structured, professional AI prompt.

Try it free โ†’

Frequently Asked Questions

No. These prompts write a curriculum you can stand in front of: outcomes, modules, lessons, exercises, a launch plan, and a sales-page outline. You still record the lessons and teach the live sessions. Claude does not replace your voice, your examples, or your grading.
Teacher prompts plan a class period, a rubric, or a substitute folder. This page plans a product you sell and teach: a finishable outline, practice, launch week, and a sales outline. Use teacher prompts for tomorrow's period. Use this page when the outline has to survive a cart open.
Yes. Lock outcomes and modules first, then fork lessons for the format. Cohort needs live unlocks. Self-paced needs onboarding and a capstone students can finish alone. YouTube needs shorter lessons and a recording plan. Do not let the sales outline add a module the other forks will not record.
It will if you ask for a finished sales page with empty proof. These prompts forbid new names, quotes, enrollment counts, and job outcomes. If a proof is not in what you pasted, the draft should say ASK ME FOR PROOF. You still record the sales video from results you can show.
Yes. Paste notes, examples, and skills you can teach live. An outline from a topic name alone is a pretty skeleton you cannot record. If a lesson needs a walkthrough you have never run, stop and do the work first, then come back to the prompt.

Prompts are the starting line. Tutorials are the finish.

A growing library of 300+ hands-on tutorials on ChatGPT, Claude, Midjourney, and 50+ AI tools. New tutorials added every week.

7-day free trial. Cancel anytime.