ResourcesCompare
LiteSurface vs. spreadsheetsWhen a sheet is enough, and when it is not.
Most teams already score ideas in a spreadsheet, and many should keep doing so. Here is where a sheet wins, and where a system that researches, cites, and checks its own arithmetic starts to pay off.
Start here
When the alternative is the better choiceReasons to stay with a scoring spreadsheet.
You have a handful of ideas
Five ideas and an afternoon do not need a pipeline. A shared sheet with agreed criteria is faster to set up and easier to change on the spot.
The evidence is already in your heads
If the team knows the market first-hand and needs no desk research, a sheet captures the judgement without extra steps.
You want total manual control
A sheet does exactly what you type. Custom formulas, odd weightings, and one-off exceptions are trivial.
Cost has to be zero
A spreadsheet costs nothing to run. LiteSurface is free during early access, but live research and evaluation spend model credits on your provider keys.
The differences
What changes, and only what you can checkFive differences, no scores for ourselves.
Each row is something you can verify: in the product, in the sample evaluation, or in the methodology.
| Topic | Scoring spreadsheet | LiteSurface |
|---|---|---|
| Who scores | Whoever fills in the sheet. A second scorer is possible if you organise it. | Two independent evaluators (OpenAI and Anthropic) score every criterion; code takes the median and keeps the disagreement visible. |
| Deal-breakers | Only if someone writes the rule, and it only works when the cell is filled in. | Six gates, five of them hard, are checked on every evaluation and override the score when they fail. |
| Evidence behind a score | Links or notes in cells, when someone adds them. Nothing checks that a link supports the number. | Scores cite atomic claims stored with a locator; citations to claims that do not exist are dropped, and coverage is measured. |
| Ideas over time | Rows and file history. Changing an idea overwrites the old version unless you copy it. | Concepts are typed, versioned genomes you can fork and merge; each evaluation is pinned to the version it scored. |
| After the decision | You write the spec yourself. | A 26-section implementation handoff, checked for completeness and consistency, with citations resolved. |
Using both
Download the scorecard template as a CSV and use it in your sheet. The criteria, weights, gates, and bands are the same ones LiteSurface applies.
Questions
Fair questionswith straight answers.
You can add concepts by hand or generate them from capabilities and evidence. Importing a spreadsheet of concepts directly is not supported today.
Yes. The default scorecard has nine weighted criteria, and each project can use its own scorecard with custom criteria, weights, and gates.
Yes. The scorecard template downloads from the home page as a CSV, with no email or sign-up required.
Try it on the ideayou are weighing right now.
Solo and Team are free during early access. Explore a sample project before you run your own.
- Solo and Team are free during early access
- Explore a sample project first
- Bring your own OpenAI and Anthropic keys