Course Creator Studio — AI Task Kit (4 Agents)

From AGENTORY — ready-to-use AI Task Kits. Drop in. Power up.

Not a prompt pack and not a course platform — a filled, faithful Course Book: your bio and promise recorded exactly as you state them, your source material reproduced exactly as you pasted it with its figures exactly as printed, your lessons built only from the sources mapped to them, your launch text asserting only recorded facts, and every deliverable gated by a named completeness check, kept current every time a fact changes. Works with the AI you already have: upload one file, tell it "Start", answer a few questions, and the Book starts filling with YOUR course, your sources, and your dates.

This is not a prompt pack, not a blank template, and not a course platform. Four specialist agents fill one Course Book with your facts — your bio exactly as you state it, your course promise exactly as you record it, your title, tone, and price exactly as you state them, your source material exactly as you paste it — build lessons only from the sources mapped to them, check each deliverable against a named completeness gate, and keep the whole Book current when a launch date is reported moved, a source is added, or a status is reported. Nothing in the Book is invented, no source figure is corrected, no conflict between your own sources is resolved, and nothing leaves it as a send or a publish. A blank template holds whatever you type, errors included, and checks nothing; a bracketed prompt holds none of your facts and enforces no rule. This kit is the other thing: the Book is built — from your material, with its claim-trace gate applied to every lesson, page, email, and reply.

No new app. No subscription. These agents work with any AI chat — ChatGPT, Gemini, Claude, Copilot, whichever you have (free versions work).

How it works (one agent at a time, once):

1. Add the agent file to your AI chat — upload the .md file, or on a phone, open the .txt version, copy it, and paste it in.
2. Type "Start".
3. Answer the agent's questions, one at a time — your bio exactly as you state it, your promise exactly as you record it, your source material pasted exactly as you have it, your outline and per-lesson source map as you state them. That's the calibration, and it's what makes the Book fit YOUR course.
4. Save the complete updated Course Book each agent hands back to you — the agent keeps it current in the chat and returns it whole, never as pieces for you to assemble — and the next agent starts from your real course and sources.

Use the agents in order:

1. Course Architect — the Book's front door, and the kit's lead agent. It opens the shared My Course Book, owns your creator record, your course record (title, promise, tone, and price as stated), the source index, the outline and per-lesson source map, and the conflict register — sources reproduced exactly as pasted, figures exactly as printed, and where two of your own sources disagree, both standing with the conflict flagged, never resolved. It runs GATE-O. It creates no Gmail draft, and it performs no course-platform act of any kind.
2. Lesson Writer — the Book's drafting bench. It drafts a lesson only from the source or sources mapped to that lesson in the Book: steps in full and in source order, figures exactly as printed, a common-mistake note only where the mapped source carries one, a check-for-understanding drawn only from that lesson's own content, and a gap list on every draft. A lesson with no mapped source is listed as source-needed and is not drafted — not from the model's own knowledge, not at any level of insistence. It runs GATE-L. It creates no Gmail draft, and it performs no course-platform act of any kind.
3. Launch Coordinator — the Book's launch desk. It builds the launch plan from your recorded dates with every countdown and interval recomputed, and assembles launch drafts — the sales page and the launch email sequence — only from recorded facts: title as stated, bio verbatim, promise verbatim and never upgraded, price as stated, dates as recorded, and any testimonial only as you pasted it, verbatim, with its attribution as you stated it. It runs GATE-P. Where a Gmail draft route verifies in a session, it may create its own launch email drafts in your own Gmail drafts, after you approve that exact text — a draft, never a send. It publishes nothing and performs no course-platform act.
4. Student Support Drafter — the Book's answer desk, on a short leash. It builds the support queue from student questions exactly as you supply them, and drafts replies resting only on recorded course content: a question the course covers is answered from the cited lesson and source; a question the course does not cover gets a draft that says so, in your recorded tone, plus a gap note to you — never an improvised answer. A question carrying a legal claim, a chargeback or refund pressed as a dispute, a threat, or a safety concern is never drafted as a reply at all: it is escalated to you by name. It runs GATE-Q. Where a Gmail draft route verifies in a session, it may create its own reply drafts in your own Gmail drafts, after you approve that exact text — a draft, never a send. It performs no course-platform act.

One Course Book, nothing invented. All four agents work from the same shared My Course Book — your calibration for all four agents, your creator and course records, your source index with every pasted source reproduced as pasted, your outline and source map, your conflict register, your lesson register with statuses as reported, your launch register with milestones as recorded and testimonials as pasted, your support register with questions as supplied and escalations logged, your open "To confirm" list, and the change log exist once, in one place. No outline, lesson, page, email, queue, or reply may introduce a fact that is not in the Book or supplied by you in that session. Anything missing is printed as "To confirm" or "not recorded" — never completed from a pattern, never given a plausible credential, figure, date, or claim, and there is no enrolment, student-count, revenue, or completion field anywhere in the Book at all.

What this kit does NOT do — stated plainly, not in fine print. This kit sends nothing and publishes nothing — no lesson, no sales page, no launch email, and no support reply is sent, published, uploaded, or enrolled by any agent, in either mode, at any level of insistence. Every outbound text it prepares is a draft or a document: the agent shows you the exact text, you approve that exact text, and you send or publish it yourself. Where a Gmail draft route verifies in a session, the Launch Coordinator or the Student Support Drafter may create its own draft in your own Gmail drafts — a draft, never a send: no send step exists in this kit. No course-platform act is performed by the agents on any platform: building the course, uploading lessons, publishing a page, and enrolling students are steps on checklists the creator executes — you do them, in your own accounts, from the kit's gated documents. It invents no student, no result, no income figure, and no testimonial — and it refuses to write them when asked, in any wording, including "as an example". It corrects no source figure — not to common practice, not to another of your sources, not by rounding or averaging — and it resolves no conflict between your sources: both sides stand, flagged, until you state a resolution as your fact, which is then recorded with the originals unchanged. It never turns your recorded promise into a guarantee — not in a headline, not in a subject line, not by implication or layout. It gives no tax, legal, or platform-policy advice and states no such rule as fact — such a question becomes a drafted question for your accountant, your lawyer, or the platform, filed unanswered.

Complete, or declared incomplete — the four completeness gates. Each of the kit's four main deliverables runs a named completeness check against your recorded Book before it is presented as finished, and the result is printed with the deliverable: GATE-O for the course outline, GATE-L for the lesson draft, GATE-P for the launch plan / sales page / email sequence, and GATE-Q for the support queue. The outline check requires the creator and course as stated, verbatim, sources as pasted with figures as printed, every conflict flagged and unresolved, the outline as stated with its counts reconciled on the page, and a source map present for every lesson or that lesson printed source-needed; the lesson check requires the overview traced to the mapped source, the steps complete from that source in its order with none added or dropped, the figures exactly as printed, the common-mistake section present as source-only or declared absent in its exact form, the check-for-understanding answerable only from the lesson itself, and the gap list present; the launch check requires every date recorded with its countdowns recomputed, the promise verbatim and never upgraded, the bio verbatim, the price as stated with no discount or anchor added, any testimonial only as pasted, verbatim and attributed, no enrolment, results, or scarcity claim anywhere, and the gaps printed; the queue check requires every supplied question present as supplied, every in-course draft traced to recorded content, every out-of-course

Source figures are the course, exactly as printed. Every figure a source prints stands exactly as printed — a hydration percentage, a feeding ratio, a schedule — never corrected to common practice, never rounded, never standardised to another source's units or wording. What the kit does compute, it recomputes with the parts shown: outline counts from the recorded outline and source map, weekdays derived from their dates, and intervals and countdowns from two recorded dates. If two of your own sources print different figures for the same item, both figures are printed and the disagreement is flagged — never silently resolved, never averaged, never explained away, and never settled by preferring the newer source. In the kit's materials, one source prints 75% and an older one prints 70% for the same item, and both stand in exactly that way (see the worked example below — a fictional example, labelled as such).

Your Book holds other people's information — the privacy stance. The Book holds your course texts and, later, student questions you supply — other people's information. The agents never publish, post, or share your Book anywhere; student detail appears outside the Book only inside the exact reply draft you approved, as that draft's purpose requires — and you send that draft yourself. Passwords, account numbers, payment details, API keys, and identity-document numbers are never recorded in the Book at all and never pasted into a chat — a secret pasted into a chat gets one answer: refuse, do not use it, change it now. Keep the Book where only the people you choose can see it, share only the part a person needs — a student who needs their reply never needs your support register — and know that anything you paste into a chat is shared with that AI provider under that provider's terms. Every agent tells you this once, plainly, at setup.

Two ways to work — your choice, per session. Run an agent in a regular AI chat (ChatGPT, Gemini, Copilot) and it works as a guide: it creates and edits the complete deliverables right in the chat — the finished documents and complete updated files themselves, never outlines, edit lists, or pieces for you to assemble — and guidance is what covers the steps inside outside tools and accounts, which you perform yourself. Run the same file on an agentic AI — Muse, Claude Code, Devin, and similar — and it can do more where the platform genuinely allows: it can maintain your Course Book as a real file under git, with a commit before the first edit and one per verified step. The route fences are exact, and they do not widen by platform: Gmail drafts — the Launch Coordinator and the Student Support Drafter only, each its own drafts only, in your own Gmail drafts, after you approve that exact text — a draft, never a send, on every platform. No course-platform act is performed by any agent in either mode — on Teachable, Thinkific, Podia, Gumroad, and Payhip alike, the platform steps are checklists you execute yourself from the gated documents, and the annex language below records what each official route allows, with its limits named. A launch-checklist mirror in Todoist is the one further designed route: Launch Coordinator only, mirror-only, after the plan passes its gate, on your approval — the Book stays the source of truth and the mirror executes nothing about the course itself. Same agents, same rules — what changes is what your platform lets them do. You pick the mode every time; guidance is always available and always complete.

What "connected" means here — stated plainly, with the limits named. Capability depends on your platform and on what you choose to connect, and the official routes are limited: Teachable — limited: an official route exists that is experimental, limited in rollout, and Claude-only at the dated checking, with lesson creation not verified — the agent drafts the curriculum and you build and publish it in Teachable. Thinkific — guide-only: no official route was verified — the agent prepares everything and you build and publish it yourself. Podia — limited: an official route exists under which lessons and emails can land as reviewable drafts, but some actions — website edits are the named example — take effect immediately and cannot be undone, and there is no video upload by that route — and still, this kit's agents perform no Podia act; those are your own acts from the kit's gated documents. Gumroad — limited: an official route exists that is Claude-only at the dated checking, under which new products start as unpublished drafts and product files must be added in Gumroad before publishing; Gumroad's unit is the product and it has no lesson or curriculum builder in the official pages checked — this kit's agents perform no Gumroad act. Payhip — no route: no official route exists for creating products or courses — the agent prepares the package and copy, and you create and publish everything yourself. Any line implying this kit uploads your lessons, publishes your course, enrols your students, sends your emails, or processes a payment would be wrong, and no agent in this kit will make that claim.

Not sure what your AI can actually do? Each agent starts by telling you which platform it is running on and what it can really do in this session — its actual tools, not the brand name, because brand is not a guarantee and plans change; the in-session check governs over every statement in this listing and in the kit's dated capability annex. The kit also includes a short, plain-language guide to AI limitations, with that dated annex listing which platforms can act and which only guide, and what each can really do with this kit's tools — so you always know where you stand.

A realistic first session: with the Course Architect you answer its 10 questions one at a time and reach the first milestone: the Book opened — its cover with you as you stated yourself, your course exactly as stated, your sources as pasted with their figures as printed, every conflict between your own sources flagged unresolved, and anything missing shown as "To confirm" — plus a gated course outline with its counts reconciled and every lesson mapped or printed source-needed, a change log opened with its first entry, and one faithful first tracing demonstrated live: a lesson you name, with its recorded figures returned exactly as printed in its mapped source, attributed to that source. That is the first session — not a written course. The lessons, the launch, and the support queue come step by step after that, from the Book you build.

The worked example: the materials walk one complete fictional example end to end — Elena, a former bakery owner teaching sourdough, opening her Book on Saturday 2027-08-07 with a course she states herself — "Sourdough From Starter to Loaf" — at $129, her bio and her promise and her tone ("Warm, plain-spoken, no hype.") recorded in her own words, and three pasted sources: her "Starter Guide" prints the main loaf's hydration at 75%, a starter feeding ratio of 1:2:2, and a feeding schedule of twice daily; her older "Early Bakery Notebook" prints 70% for the same loaf — a conflict the Book flags with both figures standing, never resolved and never averaged. Her stated outline is 4 modules and 12 lessons, 3 per module: 11 lessons mapped to a source, and exactly one — Lesson 4.3, "The Wholemeal Variation" — recorded as source-needed, and not drafted, until she supplies a source for it. Her Lesson 1.2 draft is built only from its mapped source: six source-traced steps in the source's order, the ratio 1:2:2 and the schedule twice daily exactly as printed (a request to "fix" the ratio to the one everyone else uses is refused — common practice is not a source), its common-mistake section printing "none in source — not supplied" because her source carries none, its check questions answerable only from the lesson itself, and its gaps listed. Her launch is recorded as Saturday 2027-09-18 — 42 days after the Book opened — with a sales-page draft milestone on Friday 2027-09-03 (15 days before launch) and a launch email sequence milestone on Friday 2027-09-10 (8 days before launch); her sales page carries her title, her verbatim bio, her verbatim promise — never upgraded — her stated price of $129, and her module and lesson counts, prints her delivery and format facts as "To confirm" because she recorded none, prints no enrolment figure of any kind because none exists in her Book, and prints its testimonial section as "none pasted" because she pasted none in this example. Her support queue holds two questions: one her Lesson 1.2 content answers, drafted from that content only (1:2:2, twice daily), and one her course records no content for at all — "How do I make gluten-free sourdough?" — which gets a draft that says so in her recorded tone, plus a gap note to her, and no improvised answer, so you can see every format before you build your own. Your Book gets your own material; the example is there to show the shapes. Elena is a fictional example persona created for this kit — not a customer, not a real course creator using this kit, and not a result story of any kind.

If a tax, legal, or platform-policy question comes up: if you ask how a tax should be worked out, what a term requires, or what a platform's policy allows, the agent stops, gives no answer and no reassurance, and drafts the question — resting on your recorded facts — for your accountant, your lawyer, or the platform, filed in the Book unanswered. It never sends that question and never states a rule for your jurisdiction or for any platform as fact.

What's inside:

• 4 agent files, each in .md AND .txt format (identical content — one of them always works on your device, phone included)
• My Course Book — fill it in once (agents help you), paste it into any new chat, skip the interviews; it is the kit's single source of truth: your calibration for all four agents, plus the registers the agents maintain — creator and course records, source index and conflict register, outline and source map, lesson register, launch register and milestones, support register and escalation log, disagreement and conflict flags, professional questions filed unanswered, drafts logged as drafts, the open-items list, the connected-mode record, gate results, and the change log
• Prompt Sheet — the exact phrases to say to each agent, for every job, including the approval phrase for Gmail draft creation (there is no approval phrase for a send, a publish, an upload, or an enrolment — this kit performs none of them)
• AI Limitations Disclaimer (.md + .txt) — what AI can and cannot do, in plain language, with this kit's tools as the examples and a dated capability annex
• A step-by-step README that gets you running
• Troubleshooting & FAQ — a source figure that came back corrected, a conflict that came back resolved or averaged, a promise that came back sounding stronger than you recorded it, a lesson containing a tip you never supplied, a sales page showing a figure you never recorded, a support draft answering a question your course does not cover: the known rough edges and their fixes
• Changelog — what changed in this version

Who it's for:

• A solo course creator — a practitioner, craftsperson, former operator, or professional teaching what they know — whose course lives in several places at once: source material in documents and notebooks, an outline in fragments, lessons unwritten or half-written, a launch date circled, student questions answered from memory — and who needs one Book built only from those materials, with any disagreement between their own sources shown to them instead of quietly settled
• Creators who want every lesson traceable to a source they supplied, every figure in it exactly as that source prints it, and a lesson with no source named as source-needed rather than written from general knowledge
• Creators whose facts change mid-launch — a launch date reported moved, a source added for the lesson that lacked one, a draft status reported changed — and who want the Book re-issued from the updated facts, with every countdown re-derived and every change logged, instead of documents drifting apart

Who it's NOT for:

• Anyone looking for a course platform, a hosting or enrolment service, or software that uploads, publishes, sends, or sells anything for them — this kit is instruction files for your AI, plus the Book work you direct and check
• Anyone who wants an agent to write lessons from its own knowledge of your subject, correct a source toward common practice, settle a disagreement between your sources, turn a promise into a stronger claim, or write a student story, a results claim, or an income claim — those inventions are exactly what this kit refuses; the kit records your facts, traces your sources, does the date and count arithmetic, and drafts your questions
• Anyone who wants a Book that looks complete whether or not the material is in — an incomplete record is delivered with its gaps showing, and a completeness gate that fails is stated, never waived

What you need:

• An AI chat account of your own (ChatGPT, Gemini, Claude, Copilot, or any other — free versions work for the guided work)
• Your own material — your bio exactly as you state it, your course title, promise, tone, and price exactly as you state them, your launch dates as you state them, your source material pasted exactly as you have it, and your outline and per-lesson source map as you state them; the agents use only what you supply, and infer nothing
• A private place to keep the saved Course Book — it holds your course texts and student questions you supply, and only the people you choose should be able to see it

Honest limits:

Each agent is a text instruction file for your existing AI — it is not a trained AI model, not a separate app, not a course platform, and not a human course-building service. Prompt packs for this work exist cheap, and generic course prompts are easy to find free; blank course templates and workspaces are easy to find too. What this kit adds is the Book those products never build — filled only from your supplied facts and pasted sources, with every lesson traced to its mapped source under a claim-trace gate, every source figure held exactly as printed, every conflict flagged and left standing, every deliverable passing a named completeness gate, and the whole record kept current by re-issue and change log when your facts change. What the agents can do depends on where you run them: in a chat-only AI they guide and draft — creating and editing the complete deliverables right in the chat — and you perform only the actions inside your outside tools and accounts; on an agentic AI they can act through the tools you choose to connect — always showing you the exact action first, always with your approval before anything is created, and never unattended. Connections are made through your platform's own secure sign-in and key flows — the agents never ask for, receive, or store a password or secret key, and a secret pasted into a chat gets one answer: change it now. Mode honesty holds throughout: in Connected mode the Book is kept as a file under git so a change can be rolled back; email is draft-only in both modes — where a Gmail route verifies, a draft may be created in your own Gmail drafts only, by the Launch Coordinator or the Student Support Drafter, each its own, and a draft is never reported as a send, because nothing in this kit is ever sent by an agent; and no course-platform act is performed in either mode — Teachable's official route is experimental, limited, and Claude-only with lesson creation not verified, Thinkific is guide-only, Podia's route carries actions that cannot be undone, Gumroad's route is Claude-only and product-based rather than lesson-based, and Payhip has no route — on every platform the deliverable is the gated document you place yourself. The kit's boundary holds in both modes: the Book is built only from what you supply, unknowns are printed as "To confirm", source figures stand exactly as printed and are never corrected or averaged, conflicts are flagged and never resolved by the kit, lessons are built only from their mapped sources, the recorded promise is never upgraded, no student, result, income figure, or testimonial is ever invented, statuses change only when you report them, and a confirmed change re-derives everything resting on it and flags what it conflicts with instead of rewriting the record. The agents' worked example, Elena, is a fictional example persona created for this kit to show the formats end to end — not a customer story. You check every output before you rely on it. This kit promises no outcome for your course — not a launch on any date, not enrolments, not sales, not income: what it offers is checkable, namely fidelity to your stated facts and pasted sources, recomputation with parts shown, and a completeness gate on every major deliverable.

One honest expectation: this kit is a plain-text instruction file, and it loads in any AI you already use — but the AI you run it on sets the quality. The agent's rules, questions, and output formats are fixed in the file; how deeply and consistently your AI follows them depends on your model and on how much it can hold in one session, so a weaker or free-tier model may shorten, skip, or drift where a strong current model stays exact. AI tools, models and context lengths all affect the final results — and trying different tools may help you find the best one for your needs; the same file, unchanged, often does better on a stronger model. The full disclaimer in your download explains this properly and shows how to get the best from whatever AI you have.

Price: $49 — Standalone: all 4 agents, everything included.

Starters (ONE agent, your choice of any of the 4) are $19 each — separate listings below. Starter owners complete the set with the Upgrade, $30 — the difference.