Skip to main content
Goal: you have a few versions of something (email drafts, taglines, designs) and want people to tell you which one they’d send, and why. They open one link, pick, and press Send. You get one table with every reviewer’s picks, which you can split into one row per question. This replaces the old way of doing it, where the page downloads a JSON file and the reviewer has to email it back to you.
1

Have your AI build the picker

Ask Claude, ChatGPT, or Gemini:
“Build a page that shows these 3 groups of email drafts. For each group, let the reader pick the one they’d send and say why. Add a Send button that submits the picks to Shareable with Shareable.submit, with one entry per group in a rows list. If window.Shareable isn’t there, download the picks as JSON instead.”
Or start from the example below.
2

Publish and share

Publish the page on Shareable. Because the page calls Shareable.submit, it starts collecting responses automatically. Pick who can open it in Share. Use verified email or specific people if you want each response labelled with the reviewer’s email.
3

Read the picks

Open the page in your dashboard and click Responses. Split the table on rows to get one row per reviewer per group, with the reviewer’s name copied onto each row, then download a CSV. Or ask your agent to “summarize the responses on my picker page” (it uses get_responses).

The example page

What you get back

Each reviewer sends one object like this:
In the table, that’s one row per reviewer. Split it on rows and it becomes one row per reviewer per group (reader, group, best, why). Nine reviewers and ten groups give you 90 rows you can sort, filter, or summarize: the summary then shows how often each option was picked as best.
Reviewers can change their minds. Sending again replaces their earlier picks, so you always see each reviewer’s latest answer.
Testing in the dashboard preview? It won’t send. The preview shows “Responses are only collected on the live link.” Open the published link to try it for real.