Anvil Alternative: Fill PDFs by API Without Building Templates

Sep 20, 2026·7 min read

Anvil is a well-built document platform: PDF filling, PDF generation, e-signatures, web forms, workflows, an OCR service and an MCP server, with public pricing and a free tier. If you are integrating documents into a product you own, it is a serious option and this article will not pretend otherwise.

It also has one design decision that decides whether it fits you. Its PDF filling works in two steps, in Anvil's own words: first "turn your PDFs or documents into API-fillable templates", then fill the template with structured data. Every PDF becomes a template with named fields before the first fill. For a product that fills its own ten forms a million times, that is the right model. For a team that receives a different PDF from every client, it is the whole cost.

We build FillWise, whose PDF form API takes the opposite approach, so read this as a vendor comparison. Everything about Anvil below is from its website as of September 2026.

How Anvil fills a PDF

Upload the PDF. Anvil's Document AI detects the boxes and proposes field names; you review and adjust the template.
Call the API with a JSON payload whose keys match the template's field names.
Receive the completed PDF.

Pricing, from the pricing page: a Free plan at $0 with 2 users, unlimited templates and "500 free API calls per month, $0.10 add'l", described as a UI-only plan until a card is added for API access. An AI Pack at $99 a month adds programmatic Document AI, AI schema mapping, AI-assisted box finding and field labelling, 50 pages per AI-tagged PDF and a higher rate limit. A Product Pack at $425 a month adds white labelling and signing features. Enterprise is custom.

The strengths are real: templates are reusable and precise, the field names are yours, the fill is deterministic, and the rest of the platform (signing, web forms, workflows) sits behind the same account.

Where the template model costs you

Three situations where the setup step is the problem:

Every PDF is different. A visa agency receives embassy forms, airline forms and insurance forms from dozens of sources. A law firm gets a new court or agency form per matter. Templating each one before the first fill means the first fill costs an hour of setup, and most forms are filled a handful of times.

Your data is not structured. Anvil wants JSON keyed to field names. If what you have is an email from the client, a CRM note or a passport photo, something has to turn that into JSON first: either a person or another AI step you build and pay for separately.

The mapping has to be maintained. When the embassy revises its form, the template is stale. Someone notices, usually after a bad fill.

None of this is a flaw in Anvil. It is what "template first" means.

How the FillWise API fills a PDF

One call. You send the PDF and the source data as plain text, and get the completed PDF back.

curl -sS https://fillwise.ai/api/v1/pdf/fill \
  -H "Authorization: Bearer $FILLWISE_API_KEY" \
  -H "Content-Type: application/json" \
  -d "{
    \"pdf\": \"$(base64 -w0 application.pdf)\",
    \"data\": \"Maria Petrova, born 14.03.1992 in Riga, Latvian, passport LV4482913 valid until 05.02.2031\"
  }"

No template. The engine reads the form's own labels, in whatever language, and matches the data to them. The data field takes text in any format, up to 60 000 characters, and you can attach up to four documents (photos, scans, PDFs) for the AI to read instead of, or on top of, the text. A pages parameter fills only part of a long packet.

The response carries the filled PDF as base64, filled_count, and mode: acroform when the PDF had real fields, flat when it had none and the values were drawn at label positions. Scanned PDFs without a text layer are rejected with a 422 rather than returned half-filled.

Keys are created in your account and there is no separate API plan: each successful call spends one fill from the same balance as the Chrome extension. Free is 20 fills to start and 5 a week; paid plans start at $15 for 100 fills a month, listed on the pricing section. Rate limit is 20 requests a minute per account. Full parameter and error reference is in the PDF form API guide.

Side by side

	Anvil PDF filling	FillWise PDF API
Before the first fill	Turn the PDF into a template, review field names	Nothing
Input data	JSON keyed to template fields	Plain text in any format, plus up to 4 documents
New or changed form	Update or recreate the template	Send it
Fill result	Deterministic, exact	AI-matched, review recommended
Flat PDFs without fields	Template boxes drawn by Document AI	Values placed by reading the layout, `mode: flat`
Scanned PDFs	OCR service available separately	Rejected
E-signature, web forms, workflows	Yes, same platform	No
MCP server	Yes	Yes, `fillwise-mcp` on npm
Free tier	500 API calls a month, card required for API	20 fills to start, 5 a week, no card
Paid	$0.10 per extra call; AI Pack $99/mo; Product Pack $425/mo	From $15 for 100 fills a month

Which one, decided

You own the forms. Your product generates the same contracts, applications or certificates from your own database, thousands of times, and you also need signatures. Anvil. The template is built once, the fill is exact, signing is on the same account.

You receive the forms. Clients, agencies, embassies and insurers send you PDFs you did not design, and your data arrives as text, emails and document photos. FillWise. No template, and the AI reads the document instead of you keying JSON.

You are an AI agent builder. Both expose an MCP server. The difference is the same: Anvil's tools fill templates you set up; FillWise's fill_pdf takes any PDF and free text.

Some teams will end up with both: FillWise for the long tail of incoming forms, Anvil for the handful of documents they generate and sign at volume.

What to check by hand

An AI fill is a first draft. Before a document leaves your system:

Review filled_count against the number of fields you expected; a low count on a fillable PDF usually means an XFA form, which has no standard fields.
Open the output when mode is flat; positions are computed and can sit a few points off.
Check names and dates against the source document.

Privacy

The PDF and the data in an API call go to the FillWise server and to the AI provider for that one fill, and are not stored. What the account keeps (keys, usage counts without field values) is listed on the privacy and data security page.

FAQ

Can I move my Anvil templates to FillWise? There is nothing to move. FillWise reads the PDF itself; send the original form and the data.

Is an AI fill as reliable as a template fill? No, and we say so. A template fill puts the value exactly where you mapped it. An AI fill reads the form and can misjudge a look-alike field. That is the price of skipping setup, and why the response tells you how many fields were filled.

Does FillWise sign documents? No.

Does the FillWise API fill web forms? No. Web forms are filled by the Chrome extension on the page. The API is PDF only.

What happens on a scanned PDF? A 422 with a clear code. The Chrome extension cannot fill scans either; a text layer is required.

Get started

Create a key in your FillWise account, take the last PDF a client sent you, and run the curl above with their details as plain text. If the fill count is right on that one, it will be right on the next hundred.

More guides

AI Google Forms Filler: Real Data and Test Data in One Click — How to fill a Google Form with AI from pasted text or a document, or with realistic test data for a form you are building. What fills, what still needs a click, and what it costs.
Chrome Autofill vs FillWise: When the Built-in Filler Is Enough — Chrome already fills names, addresses, cards, identity documents and travel details. Where it stops, what an AI form filler adds, and how to decide which one a form needs.
DocSpring Alternative: Fill PDFs by API Without a Template Editor — DocSpring fills PDFs by API after you set up each template in its editor. FillWise takes the PDF and plain text in one call. When each fits, with prices from both sites.
How Event Organizers Fill Registration Forms Faster with AI — A practical guide to filling event and conference registration forms from a participant list, including multi-attendee forms, with the real workflow and honest notes on limits.
How Finance Teams Fill Banking, Tax and KYC Forms with AI — A practical guide to filling banking portals, tax systems and KYC questionnaires from client records, statements and invoices, and what to double-check by hand.
How to Fill a Flat PDF Form (No Fillable Fields) — Flat PDFs have no interactive fields, so most tools can't fill them. Here's why they exist, the options for filling one, and how AI can place text on the right lines automatically.
How Insurance Agents Fill Policy Applications Faster with AI — A practical guide to filling policy applications and ACORD forms from photos of client documents, with the real workflow, how PDF forms are handled, and honest notes on limits.
How Legal Teams Fill Court Forms and Client Intake with AI — A practical guide to filling court e-filing forms, client intake and registry paperwork from case data or scanned documents, plus what to verify before filing.
Magical Alternative for Form Filling in 2026 — Magical has moved to healthcare AI agents and hides its pricing. If you used it to fill forms from client data, here is what changed, what to look for in a replacement, and how FillWise compares.
How to Fill PDF Forms Automatically with AI — A practical guide to auto-filling fillable PDF forms. How to tell if your PDF can be filled, the exact workflow, how checkboxes and dropdowns are handled, and why the document never has to leave your browser.
PDF Form API: Fill Any PDF Form from Your Code — A practical guide to the FillWise PDF form API. One HTTP call, the PDF and your data in, the completed PDF out. Request fields, error codes, limits, and curl, Python and Node examples.
How Recruiters Fill ATS and Job Board Forms Faster with AI — A practical guide to moving candidate data from resumes and LinkedIn profiles into ATS forms, job boards and onboarding paperwork, with the real workflow and honest notes on what it does and doesn't do.
How Sales Teams Move CRM Data into Web Forms Faster with AI — A practical guide to filling lead qualification, outreach and proposal forms from your CRM data without re-typing, with the real workflow and honest notes on what it does and doesn't do.
Simplify vs FillWise: Which Form Filler Do You Need? — Simplify Copilot autofills job applications for free. FillWise fills any form on any site, plus PDFs, from text or a document. An honest comparison of where each one is the right pick.
How to Fill Web Forms from Spreadsheet Data with AI — A practical guide to moving spreadsheet rows into web forms without copying cell by cell, with the real workflow and an honest explanation of what the tool does and doesn't accept.
How Travel Agents Speed Up Visa Applications with AI Autofill — A practical guide to filling visa applications, DS-160 forms and booking portals from a passport photo or CRM data, with the exact workflow, what to double-check before submitting, and honest notes on limits.
How AI Web Form Autofill Works, and When to Use It — An overview of filling web forms automatically with AI, from any text, photo or PDF source, how field matching works across field types and languages, and honest notes on its limits.