You're viewing a live demo with sample data. The real dashboard runs on live GitHub and business data behind sign-in.

Ideation Studio

Paste a raw idea. The studio works out which of eighteen lenses it actually deserves, then builds those prompts with your own portfolio already in them. Copy them into Claude or ChatGPT — the riffing happens there, the context comes from here.

Drop a PDF, .md, or .txt here, or
Read in your browser — the file is never uploaded anywhere.
Routing keys off these. A misread stage selects the wrong half of the library.
stage
type
dependency
novelty
routed automatically · 11 qualified
Constraints (edit if these have changed)
loading your registers…
1. Have I already killed this?
Here are ideas I previously considered and rejected, with reasons:
(kill register unavailable — could not read strategy/IDEAS.md)

Here is a new idea:
(paste your idea above)

Is this new idea the same as, or a close variant of, anything I already killed?
Be strict — a surface difference (different industry, different name, different
tech) does not make it new if the killing reason still applies.

If it IS a repeat: name the previous idea, quote the original kill reason, and
tell me whether that reason still holds. If it no longer holds, say exactly what
changed — new evidence, not new enthusiasm.

If it is genuinely new: say so in one line and stop. Do not pad.

A good answer looks like: Either a named repeat with the original kill reason quoted, or one line saying it's genuinely new.

2. Bottoms-up demand math
(paste your idea above)

Do NOT give me a TAM. Do not cite an industry report. Build the number from the
bottom up, and show every step so I can attack the assumptions individually.

1. Define the buyer so narrowly I could build a list — industry, size, country,
   job title. Not a segment. A list I could actually assemble this week.
2. How many of them exist within my realistic reach? Say where that count comes
   from. If you are estimating, say "estimated" and show the reasoning.
3. What fraction plausibly has this problem badly enough to spend money on it this
   year? Justify the fraction — do not reach for 10% because it sounds modest.
4. What is a realistic conversion rate from "I contact them" to "they pay," for a
   solo operator with no brand recognition in this space? Be pessimistic.
5. Multiply it out. Annual revenue at year one, and at a steady state I could
   actually service alone.

Then the only question that matters: is the year-one number worth a quarter of my
time? Compare it to just billing those hours at consulting rates. If consulting
wins, say so directly.

A good answer looks like: A number small enough to be uncomfortable. Bottoms-up almost always lands an order of magnitude below the vibe.

3. Portfolio compounding check
I am a solo operator. Reuse is my only structural advantage — anything I build
from scratch competes with people who have teams.

Here is what I already own:
(capability register unavailable)

Here is the idea:
(paste your idea above)

Answer three things, concretely:

1. What percentage of this already exists in what I own? Name the specific
   assets and what they'd cover. If the honest answer is "almost none," say so
   plainly — do not manufacture a connection to be encouraging.

2. What is genuinely new work? Be specific about the pieces, not the categories.

3. Is there a nearby version of this idea that would reuse MORE of what I own
   while solving a similarly valuable problem? If yes, describe it. This is the
   most useful thing you can tell me.

A good answer looks like: A percentage you can argue with — and, most valuable, a nearby variant that reuses more.

4. Competitive teardown

Inside Claude Code this runs as /market-competitors, which does live research instead of working from model memory — the difference between four verified competitors and eight where half are invented. The prompt below is the fallback for a chat surface with no tools.

(paste your idea above)

Map the competitive landscape. Name real products, not categories.

- Five to eight named alternatives, including the non-software ones: an agency, a
  freelancer, an in-house hire, a spreadsheet, doing nothing. Doing nothing is the
  market leader in most categories and must appear on this list.
- For each: who they serve, roughly what they charge, and the one thing they are
  clearly good at.
- Where is the gap? Not "they're expensive and clunky" — that is what every founder
  says and it is usually wrong. A real gap is a buyer segment they have
  structurally chosen not to serve, or a job they cannot do without breaking their
  own model.
- Which of them is most likely to add this as a feature and erase my position? How
  fast could they?
- If I entered, which one am I actually taking money from? If the answer is
  "nobody, we'd expand the market," treat that as a warning and explain why.

Flag anything you are unsure about — I would rather have four verified competitors
than eight where half are hallucinated. Say explicitly which names you are
confident are real and current.

A good answer looks like: The honest gap, plus an explicit confidence marker on each named competitor.

Routing, ranking, and slot-filling run here. Classification does not — reading an idea and deciding its stage is judgment, and there is no model on this page. The reader above proposes signals from keywords and tells you what it could not read. Selection is capped at four on purpose: the library’s own rule is two or three per session, because each lens produces a page of material to think about.