A survey tool with the awkward bits done.

How a survey moves through the product

You build in a drag-and-drop canvas with a grouped palette. Questions can carry an image, an inline heading and a description line, and choice options render as bento tiles by default — uniform-height cards that reflow to one per row when a label runs past about 110 characters — or as a plain list if you prefer. Options can carry their own image and description, and can be marked fixed so they stay put when answer randomisation shuffles the rest. Randomisation is seeded per page load, so going back a page does not reshuffle the options under someone's cursor.

Then you decide how it is presented. There are three layouts: classic, one scrolling page; conversational, one question at a time, where one-tap questions advance by themselves; and paged, a section per screen. Validation is per question — required, minimum and maximum length with a live counter, numeric range, email, phone, URL, UK postcode, and custom regex with your own error message.

Collection is where long surveys are won or lost. Save & continue issues a resume link so a half-finished response is not a lost one. Hidden fields read values from the query string and store them with the response, which is how you tell the mailout apart from the poster QR code. A close date and a response cap can both be set, and they are enforced by a database trigger rather than by the interface — responses insert straight from the browser, so the limit has to live where it cannot be bypassed. When a survey closes, respondents see a message you wrote rather than an error. Embed mode drops the whole thing into another page.

Finally, results. Charts per question, counts, completion, individual responses, and CSV or per-response PDF when you want the data somewhere else.

Who it is built for

It is built for longer, more structured surveys than a quick feedback form, which is the shape most public-sector and third-sector work takes: matrix grids for standard question sets, ranking for prioritisation exercises, scoring with bands for self-assessments, address capture with a postcode check, and per-survey partner logos for work run jointly with another organisation. There are pages on how that plays out for local government, charities and non-profits, education and healthcare, and for jobs rather than sectors, market research and HR and people teams.

Accounts, organisations and data

An account can belong to several organisations. Roles are owner, admin and member, and the last remaining owner is protected from removal or demotion at database level, so nobody can accidentally lock a workspace out of itself. Row-level security applies throughout and organisation isolation is pinned server-side rather than taken on trust from the client.

Everything runs on Supabase in London (eu-west-2): survey content and responses stay in the UK, encrypted in transit with TLS and at rest with AES-256. We publish a full sub-processor list and will answer a DPIA. What we will not do is imply certifications we do not hold — there is no ISO 27001 certificate and no completed SOC 2 audit, and the security page describes the actual controls instead.

Co-branding is per survey

Partner and collaborator logos attach to an individual survey rather than to the account. One team can therefore run a consultation badged for one council this month and a different one the next, without touching their own brand settings. It is deliberately kept separate from the organisation's brand kit — see co-branding.

What it does not do

A features list is only useful if the absences are in it too. None of the following exist today:

If one of those is a hard requirement, this is not your tool yet, and it is cheaper for both of us to find that out on a web page than three weeks into a procurement. Details of what is settled and what is not are on the pricing page, and the comparison pages set the same list against what other tools publicly document.