Deooch Forms
Build a form by describing it. Share a link. Watch the responses land.
Deooch Forms (forms.deooch.com) is a form builder with an MCP server bolted on. You can drag fields around in the visual studio like any other form tool — or skip that entirely and ask Claude, ChatGPT, or any MCP-capable assistant to build the form for you, then read the responses back the same way.
Respondents never sign in. They open a public /f/<slug> link, fill it in, and submit.
What it does
- Form studio — 13 field types, sections, validation, required fields, live preview.
- Public forms — every form gets a shareable
/f/<slug>link. No account needed to answer. - Submissions — responses appear in the dashboard as they arrive, exportable to CSV/XLSX.
- Email notifications — get pinged when someone submits.
- MCP server — six tools (
create_form,update_form,get_form,list_forms,list_submissions,get_submission) so an AI assistant can do all of the above for you. Published to the MCP Registry asio.github.musaib001/deooch-forms; setup walkthrough at /connect. - Workspaces — invite members, share forms, per-role access.
Architecture

- Creator flow — sign in with Google via Supabase Auth → build forms in the Next.js/React Form Studio → forms are saved to Supabase Postgres.
- Data layer — Supabase Postgres is the source of truth (forms, fields, tokens, RLS policies).
- Respondent flow — anyone with the link opens the public
/f/[slug]form (no login) and submits, which writes straight to Supabase. - Back to the creator — submissions show up live in the dashboard, exportable as CSV/XLSX.
- Automation — an MCP server exposes forms and submissions over OAuth so AI agents can create forms or read responses programmatically.
Stack: Next.js 16, React 19, TypeScript, Tailwind CSS, Supabase (Postgres + Auth), deployed on Vercel.
Field types
Short text · Long text · Email · Phone · Address · Number · Dropdown · Single choice · Multiple choice · Date · File link · File upload · Section heading
Sample forms
Things people actually build with it. Each one is a prompt you can paste into a connected AI assistant — it comes back with a live link.
Employee onboarding
Create an employee onboarding form with a Personal Information section (full name, email, phone, home address), an Employment section (start date, department dropdown, manager), and a document upload for ID.
Uses: section headings, short text, email, phone, address, date, dropdown, file upload.
Customer feedback
Create a customer feedback form: how satisfied are you (1–5 single choice), what did you use us for (multiple choice), and a comment box.
Uses: single choice, multiple choice, long text.
Event registration
Create an event registration form with name, email, which sessions they're attending (multiple choice), dietary requirements (dropdown), and number of guests.
Uses: short text, email, multiple choice, dropdown, number.
Job application
Create a job application form with name, email, phone, LinkedIn URL, a resume upload, and a "why do you want this role" long answer.
Uses: short text, email, phone, file upload, long text.
Bug report
Create an internal bug report form: title, severity dropdown, steps to reproduce, expected vs actual behaviour, and a screenshot upload.
Uses: short text, dropdown, long text, file upload.
Connecting an AI assistant
Add this as a remote MCP connector in your assistant's settings:
https://forms.deooch.com/api/mcp
It signs you in through the browser — no token to copy or store — and the connection
acts as you, with your role and plan. Clients that read the MCP Registry can find it
by name as io.github.musaib001/deooch-forms instead. Full walkthrough, including
ChatGPT web setup and the errors people actually hit, is at
forms.deooch.com/connect.
Development
npm install
npm run dev # http://localhost:3000
npm test # vitest
npm run lint
Layout
src/app/ routes — (portal) dashboard, /f/[slug] public forms, /api/mcp
src/components/ builder/ (form studio), forms/, submissions/, marketing/
src/lib/ forms/ (schema + validation), mcp/, auth/, email/, export/
supabase/migrations database schema and RLS policies
Read AGENTS.md before writing code — this repo tracks a Next.js version whose
conventions differ from what most models were trained on.