Inspiration

To convert a PDF into a signable form, you have to retype every field. We wanted it to be a single sentence to your browser:

"Make this a signable form."

What it does

Two WebMCP tools on formesign.com:

create_form turns your agent's submitted document into a live signable form within your account, and drops you into the signing preview.

edit_full_form allows editing in natural language and seeing the changes as they appear on the form.

How we built it

Form generation: our document-to-form model, fine-tuned on OpenAI models, converts your agent's HTML into the full form: fields, agreement text, signature blocks.

WebMCP layer: exposes only the tools that you need on the page you are at, always acting as the logged-in user.

Challenges we ran into

Sign in. Forms are in your account, but our homepage has no auth state at all. Our first attempt was a separate get_login_status tool that the agent had to call before anything else. It felt clumsy, so we scrapped it: now create_form checks the sign-in itself, and if you are not logged in, it just redirects the tab to Google sign-in and the agent picks up from where it left off.

Accomplishments that we're proud of

Onboarding has gone from 15-20 minutes of uploading, reviewing and clicking to under a minute of conversation. The user only does about 5% of the work: describe, sign in, review.

What we learned

With WebMCP, the page dictates which tools exist. On the homepage we offer to create a form; within the editor we offer to edit it. The agent's world always matches your screen.

What's next for Formesign WebMCP: Document to Signature Form

Reporting and onboarding tools. Ask "who has signed so far?" in your browser, or go from landing page to first signed document in one conversation.

Built With

Share this project:

Updates

Submission history