Playbook
From a signal in the database to a product people pay for.
Finding a good idea is the start, not the finish. This is the exact path members take — search, plan, execute, and scale — and the discipline that makes each step pay off.
- 01
Search
One idea with real demand behind it — and the evidence to back it.
- 02
Plan
A shortlist, a co-founder or a solo plan, and a paste-ready build prompt.
- 03
Execute
A working v1 in front of real users, with the build state written down.
- 04
Scale
Your first distribution, a roadmap grounded in real use, and the next idea.
Search
Pick the idea you can reach customers for, not the one that sounds biggest.
You'll end with
One idea with real demand behind it — and the evidence to back it.
- 01.1
Filter to what fits you
Start from category, customer type, and build difficulty. Sort by opportunity score. The best idea for you is one you can build and one whose customers you can actually reach.
- 01.2
Read the evidence, not the pitch
Open the sources, the competitors, and their pricing. Repeated, specific complaints plus a business already charging for a version of it is real demand. A clever summary with nothing under it is not.
- 01.3
Look for a wedge
The strongest ideas name a specific underserved angle and an audience that already gathers somewhere. That is where a small team wins against a crowded market.
- 01.4
Watch for red flags
Vague pain, no sign anyone pays, a build that needs months before a first version, or a saturated space with no angle. When the evidence is thin, move on.
Plan
Commit to two or three, then decide who builds and how small the first cut is.
You'll end with
A shortlist, a co-founder or a solo plan, and a paste-ready build prompt.
- 02.1
Save your shortlist
Save the two or three ideas that fit your skills and reach. Your saved list is what you actually commit to, and it is what the MCP and co-founder matching work from.
- 02.2
Co-founder or solo
If you want a partner, Eureka matches you with members who saved the same idea. Someone who already believes in it beats someone you have to convince. If you build solo, plan for the parts you are weakest at.
- 02.3
Scope the smallest test
Read the MVP summary and build plan, then cut everything that is not the core loop. The goal of v1 is to prove the wedge, not to be complete.
- 02.4
Turn it into a build prompt
Use the idea to generate a paste-ready Claude Code build prompt. Now the plan is concrete before you write a line of code.
Execute
Ship the core loop first, put it in front of ten real users, then decide what is next.
You'll end with
A working v1 in front of real users, with the build state written down.
- 03.1
Set up the Eureka MCP
One command connects the database to Claude Code and Claude Desktop over HTTP. Now Claude can read any idea, your saved list, and your workspace directly from your editor.
- 03.2
Open a workspace
Each idea gets one shared command center: tasks, notes, decisions, GitHub activity, and build sessions in one place. It is where you and a co-founder stay aligned.
- 03.3
Build with Claude Code
Hand it the build prompt to scaffold, then work task by task. Claude Code logs each session into the workspace, so the state of the build is always visible.
- 03.4
Track progress honestly
Every session records what shipped, what is blocked, and what is next. When something stalls, it is written down, not lost in a chat history.
Scale
Let evidence, not opinion, decide what you build next.
You'll end with
Your first distribution, a roadmap grounded in real use, and the next idea.
- 04.1
Go where the users already are
The sources named the communities and channels that surfaced the pain. Those are your first ten conversations and your first distribution.
- 04.2
Keep decisions visible
As the team grows, the workspace keeps decisions and blockers in one history. New teammates and Claude Code can read the full context instead of asking around.
- 04.3
Pull adjacent ideas
As you learn what your users actually need, search the database for adjacent openings. The same research that found the first idea feeds the roadmap.
- 04.4
Hold the same bar
The discipline that found a validated idea is the discipline that keeps it alive. Ship, measure, and let the evidence pick the next move.
Provenance
Where the ideas come from
Every idea starts from public, terms-aware sources, then gets validated against comparable businesses that already make money. You can open the evidence on any idea.
Signals are mined at scale, filtered hard, then read one by one by a person before an idea is published. If the demand is not real, it never makes the database.
Data ethics
Handled with care
We treat public sources with the same three principles on every signal we store.
Official APIs first
We use official, rate-limited APIs and public pages. We do not scrape private or protected content, and we honor source terms.
Attributed and minimal
We store short, fair-use excerpts with a link back to the source. Author handles are hashed. No personal data is collected.
Deletion-aware
When upstream content is removed, our deletion-sync removes the excerpt and any identifying fields. We do not retain deleted content.
The idea is the easy part. Knowing it's worth building is the hard part.
The Eureka Database does the validation work most founders skip: evidence the problem is real, proof that people pay for it, and a clear plan to build. Two ideas are open to read in full, no account needed.