PDF.co API + MCP by PDF.co (Artifex Software)

# Fix list: PDF.co API + MCP

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/pdf-co, the October 2026 research run, assessed 1 October 2026. Grade E, 38.2 out of 100.

This is everything the published grade says the listing lacks, the biggest possible gain to the total first. It comes from the reason given for each score, the checklist each category was scored against (https://www.anchorterminal.com/benchmark/#checklist), the provenance checks, the deductions, what we couldn't check and what the review panel asked for. A fix counts at the next check, once it's public.

For a coding agent working on PDF.co API + MCP: work through the items below in the product, its docs and its public pages. Each category gives the reason for its score, with the points each checklist item earned, and the checklist itself, so the gap is the items that earned less than their points. Change the product, not the wording, and keep a note of what you changed and where it's published.

## 1. Reliability, 20 out of 100, up to 16 more on the total

Why it scored 20: status.pdf.co answers 302 to betterstack.com/uptime, Better Stack's own product page, so there's no status page (0) and no readable incident history (5). The June 2026 changelog moved rate-limit windows from per second to per minute, but no page we read gives the numbers (0). The response-codes page and the OpenAPI document define 429 as too many requests, with no Retry-After or backoff guidance. We gave 5 of 15 for the documented code and the async job and webhook-retry route. No SLA in the terms of 24 March 2026, which reserve the right to change or stop the service without notice (0). GA (10).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-reliability):

Hosted APIs, MCP servers, models and platforms.

- 20, a public status page with component history (Statuspage, Instatus, BetterStack or the vendor's own).
- 0 to 30, the incident record for the last 90 days on that page. 30 for a clean record or trivial incidents only, 20 for minor incidents only, 10 for one major outage (an hour or more of a core API down, or errors across the board), 0 for several. 5 when there's no history we could read, and the note says so.
- 15, rate limits documented with numbers.
- 15, documented 429 or overload handling (Retry-After, backoff guidance), and idempotency keys or safe-retry guidance where writes are involved.
- 10, an SLA published for any paid tier.
- 10, the surface agents use is generally available, not beta or preview.

Local packages, SDKs, frameworks and stdio MCP servers.

- 20, installs from an official package with supported runtimes stated.
- 25, a public CI and test suite, passing on the default branch.
- 0 to 25, open crash or regression issues relative to activity (25 for few and handled, 0 for many, old and unanswered).
- 15, semver discipline and breaking changes called out in a changelog.
- 15, version 1.0 or later, or declared stable.

Protocols are read from their reference implementations, the public facilitators or servers, spec stability and test vectors.

## 2. Security & auth, 23 out of 100, up to 13.5 more on the total

Why it scored 23: One API key per account in the `x-api-key` header, header only per the OpenAPI document. We found no scopes, no multiple keys and no rotation guide (10). Every MCP tool also takes `api_key` as an argument, which invites the key into the model's context. No read-only mode, no annotations, and `upload_file` reads any absolute local path and uploads it to PDF.co. The SSE and HTTP transports bind 0.0.0.0 with no authentication (0). PDF text, tables and invoice fields come back as untrusted content with no injection guidance (0). Each response reports credits used and remaining, which is a per-call cost trail. We found no audit log (5). The security page claims SOC 2 Type II and cites AWS's ISO 27001, PCI and FISMA, which are the data centre's. security.txt is a 404 and we found no disclosure policy or bounty (8).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-security):

- 0 to 30, the credential model. 30 for OAuth 2.1 with scopes, or scoped and revocable keys with rotation. 20 for plain revocable API keys. 10 for one all-powerful key. 10 off when a secret can travel in a URL query string as a documented option.
- 0 to 20, read-only or least-privilege modes, and confirmation or approval for destructive actions.
- 0 to 15, prompt-injection posture where the tool returns untrusted content (documented mitigations or guidance). A tool that returns no untrusted content gets 10.
- 0 to 15, audit logs or per-call visibility for the operator.
- 0 to 20, a security programme. security.txt or a disclosure policy, a bug bounty, SOC 2 or ISO 27001, advisories handled in public.

Models are read for retention, whether API data trains models (and whether that's off by default), zero-retention options and certifications. Frameworks for telemetry defaults, approval hooks, guardrails and sandboxing.

## 3. Payments & pricing, 20 out of 100, up to 10 more on the total

Why it scored 20: No x402, MPP or L402 (0). Plan prices are public and the credit cost of each endpoint is published per page or per call, so per-unit cost is readable without a login (20). A sign-up trial exists per the 30 September check, but neither the pricing page nor the terms state its size or whether a card is needed, so we scored it absent (0). A person signs up in a browser to get a key (0).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-payments):

The published rubric, also on the [x402 page](https://www.anchorterminal.com/x402/).

- 40, a machine payment protocol (x402, MPP or L402) on the tool's own endpoints. 10 to 30 when it covers only some endpoints or only goes through a third party, and the note says which.
- 20, per-call or per-unit pricing published without a login. 10 for public plan-only pricing, 0 for "contact sales" or prices behind a login.
- 20, a free tier or trial that doesn't need a card.
- 20, autonomous onboarding, meaning an agent can get access without a person signing up in a browser (keyless use, x402, a programmatic key API).

Payment platforms and agent wallets rarely charge for their own API over a machine protocol, so the first line has steps for them, and the highest one that applies counts. 40 when x402, MPP or L402 runs on all their own endpoints, 30 when it runs on part of their own API, 25 when their merchants can accept one, 20 for running a facilitator, 15 for paying as a buyer, and 0 when the only protocol is their own. Merchant acceptance sits above a facilitator because the platform's own customers can charge agents through it, while a facilitator settles for sellers who wire up the protocol themselves. The counter-argument (a facilitator does more for the protocol as a whole) has a point. Each note says which step applied.

Open-source software you run yourself is scored on its hosted or paid option if it has one. A free, self-hosted package with nothing to buy gets 20, 20 and 20 for the last three lines, and 0 to 40 for the first only if it ships a payment protocol.

## 4. Agent ergonomics, 49 out of 100, up to 8.3 more on the total

Why it scored 49: 38 MCP tools from about 78 KB of source with no toolset switch, which is the more-than-30 band (5). The API itself lets a call limit pages and a rectangle and returns output as a link rather than inline, so responses size well (15). We averaged the two to 10 because the listing is the API and the MCP server together. Page ranges, `rect`, `inline` and named outputs, but nothing paginates large text results (14). Documented error codes with messages, including 402 for credits, but the MCP server labels every API answer `status: working`, error or not, and returns raw exception text on failure (12). No tool annotations, no `readOnlyHint` or `destructiveHint`, no idempotency keys, and a retried conversion is a second charged job (5). Only `url` is required and the MCP server runs async by default with a `wait_job_completion` helper. We didn't confirm official SDK packages beyond code samples (8).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-ergonomics):

- 0 to 25, context cost. For MCP, the number and size of the tool definitions (25 for ten or fewer compact tools, 15 for 11 to 30, 5 for more than 30, plus up to 10 back for toolsets, dynamic loading or read-only subsets). For APIs, whether responses can be sized (field selection, limits, summaries).
- 20, pagination, filtering and output-size controls.
- 20, actionable, documented error responses, codes and messages an agent can recover from.
- 20, idempotency or safe retries, and for MCP the `readOnlyHint` and `destructiveHint` annotations.
- 15, sensible defaults, few required parameters, and official SDKs in at least two languages.

Models are read for tool use, structured output, prompt caching, context length, batch and SDKs. Frameworks for how much code and how many defaults a tool-calling agent with MCP needs.

## 5. Maintenance & community, 34 out of 100, up to 5.8 more on the total

Why it scored 34: MCP server 0.0.9 on 2026-07-28, 65 days before this check (20). Two releases in the last 90 days, 0.0.8 on 2026-07-08 and 0.0.9, and both only repoint documentation links. The API changelog's last entry is June 2026, outside the window (0). No open issues on the MCP repository, which also means nothing to judge replies by. Support runs through the dashboard and Zendesk or Intercom per the privacy policy (12). Not in the official MCP registry. The only `pdfco` entry there is a third-party wrapper from mcparmory (0). The only CI workflow publishes releases. No tests, and the lock file dates from June 2025 (2).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-maintenance):

- 0 to 30, time since the last release, or the last published model or API change for a closed service. 30 within 30 days, 20 within 90, 10 within 180, 0 older.
- 20, at least three releases or dated changelog entries in the last 90 days.
- 0 to 25, responsiveness. Issues and pull requests answered on GitHub (the open issues and how recent the replies are). For closed services, a public changelog and a support or community channel that answers, 0 to 15.
- 15, presence in the official MCP registry under a verified namespace (MCP servers), or current official SDKs (APIs and models).
- 10, package health, current dependencies and CI.

Models are read for deprecation notice periods and model churn rather than release counts.

## 6. Schema & documentation, 74 out of 100, up to 4.2 more on the total

Why it scored 74: OpenAPI 3.0.1 at developer.pdf.co/openapi.json, header-only `x-api-key` scheme, and the 38 MCP tools carry typed JSON Schema from Pydantic (25). llms.txt and Markdown twins of every docs page (10). Tool docstrings state purpose and link the endpoint reference, and the URL field tells the model to use `upload_file` for local files, but none says when not to use a tool (10). No enums anywhere, since the server dropped `Literal` types for Gemini compatibility in May 2025. Page ranges, paper sizes and `line_grouping` are free strings, and the annotation and field arrays are `List[Any]` (6). The OpenAPI document describes errors 400, 401, 402, 403, 429 and 441 to 454 with a structured body, and every endpoint has samples (13). A monthly API changelog, but its last entry is June 2026, and the MCP server has tags and no changelog (10).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-schema):

APIs and MCP servers.

- 25, a machine-readable contract (a public OpenAPI file or similar; for MCP, typed JSON Schema inputs on every tool).
- 10, llms.txt or Markdown docs served for agents.
- 0 to 20, descriptions that say what a tool is for, when to use it and when not to, read from the tool definitions in the source or the API reference.
- 0 to 15, typed inputs with enums, constraints and required fields, and no free-form JSON blobs.
- 0 to 15, examples and documented error responses.
- 15, versioning and a public changelog.

Models are read from the API reference, the OpenAPI file, llms.txt, the structured-output and tool-use docs and the model cards. Frameworks from docs a model can follow, typed interfaces, examples and the API reference.

## 7. Transparency & trust, 54 out of 100, up to 4 more on the total

Made of editorial 42, provenance 65.

Why it scored 54: The MCP server is MIT. The API is closed with dated terms, so between the two bands (20). The privacy policy (25 August 2026) names Artifex Software, Inc. and says processing happens in the US, but states no document retention. The security page says output links expire after one hour by default, configurable from one minute to 24 days, and no DPA is mentioned anywhere we looked (12). No deprecation policy, and the terms allow changes without notice (0). AWS us-west-2 named as the host, and the privacy policy names its analytics and support vendors, but there's no subprocessor list (10).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-transparency):

- 0 to 30, source availability and licence clarity. 30 for open source under an OSI licence, 15 for closed with clear terms, 0 for unclear terms.
- 0 to 30, data handling and retention statements that agree with each other (privacy policy, DPA, retention periods, subprocessors).
- 0 to 20, a deprecation policy or notices with dates.
- 0 to 20, telemetry disclosed with an opt-out (local software), or subprocessors and data locations disclosed (hosted).

The other half of Transparency and trust is the provenance score, computed from checked facts (below). The category score is the mean of the two.

Provenance checks not met in full (half of this category, computed from checked facts):

- Domain age: pdf.co, no registry record we could read (0 of 15)
- Status page: not found (0 of 10)
- security.txt: not found (0 of 10)

## What we couldn't check

What we couldn't read counted as absent. Publishing it on a page a plain HTTP fetch can read (not only in a browser) lets the next check count it.

- unchecked: the trial's credit amount and whether it needs a card. Neither the pricing page nor the terms state it
- unchecked: rate-limit numbers per plan after the June 2026 change to per-minute windows
- unchecked: whether failed calls consume credits
- unchecked: whether PDF.co ships official SDK packages or only code samples
- The pricing page read showed $8.99 a month for Basic, which matches the listing's yearly-billed figure. We couldn't see the monthly-billed toggle, so the listing's $9.99 stands per the 30 September check

## Weaknesses

- No status page. status.pdf.co redirects to betterstack.com/uptime
- Rate limits moved to per-minute windows in June 2026 but the numbers aren't published, and 429 carries no documented Retry-After
- No SLA, no security.txt, and terms that allow changes without notice
- MCP server has no tool annotations, no read-only mode and no tests, and `upload_file` reads any absolute path
- Trial size and card requirement aren't stated anywhere we looked

## What costs an agent a turn today

The notes we give agents before they call it. Each one is a workaround an agent shouldn't need.

- Upload local files with `upload_file` or the presigned upload first, then pass the returned URL
- Read the `content` field, not `status`. The MCP server says `working` even when the API returned an error
- Call `wait_job_completion` after each job. Every MCP call runs async and returns a job id
- Download results within the hour. Output links expire after 60 minutes by default
- Keep the key in `X_API_KEY` and leave the `api_key` argument empty

## What the review panel asked for

- Publish trial size
- State failed-call billing
- honest status field
- restore enum types

## When it's done

Send what changed and where it's published as a dispute (https://www.anchorterminal.com/builders/#disputes, or `POST https://www.anchorterminal.com/api/v1/contact` with `"kind": "dispute"`). Disputes are answered in public, and the listing is checked again by the same checklist. Paying for an audit or a listing claim changes nothing here.