LandingAI Agentic Document Extraction

# Fix list: LandingAI Agentic Document Extraction

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/landingai-agentic-document-extraction, the October 2026 research run, assessed 9 October 2026. Grade B, 69 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 LandingAI Agentic Document Extraction: 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. Security & auth, 43 out of 100, up to 10 more on the total

Why it scored 43: Bearer key in a header, never in the URL. On the Explore plan there is one key that cannot be deleted or revoked, and a leaked key is replaced by emailing support. Team and Enterprise create several named keys and revoke them. No scopes on any plan (15 of 30, between the one-key and revocable-key lines). No read-only or restricted keys. Admin and Developer roles limit who may revoke keys, and an IP allowlist and SSO are Enterprise only (5 of 20). The API returns text from untrusted documents, and no prompt-injection guidance was found in the reviewed documentation (0 of 15). The Usage page logs every job with timestamp, API key, endpoint and credits, searchable and downloadable as CSV (10 of 15). The security page claims SOC 2 Type II, HIPAA with a BAA, third-party penetration tests and immutable audit logs. The libraries' SECURITY.md takes reports through GitHub private reporting. No security.txt and no bug bounty were found, and the trust centre could not be read (13 of 20).

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.

## 2. Payments & pricing, 40 out of 100, up to 7.5 more on the total

Why it scored 40: No x402, MPP or L402 (0). Credit price and per-page and per-character credit rates are public (20). 1,000 free credits on signup, and the pricing FAQ says no card is needed (20). A person signs up in a browser and copies the key from the web app (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.

## 3. Agent ergonomics, 75 out of 100, up to 4.1 more on the total

Why it scored 75: Graded on the HTTP API. Responses can be reduced with `options.pages`, block options and `atomic_grounding` set to false, and jobs can deliver output to the caller's storage with `output_save_url`. No field selection (18 of 25). Job lists take `page`, `pageSize` up to 100 and a status filter (16 of 20). Errors are documented by status code with the message text, the cause and the fix, and failed jobs carry a `code` and `message` (18 of 20). No idempotency key. Timed-out synchronous requests are not billed, the libraries retry connection errors, 408, 409, 429 and 5xx twice, and the CLI serves identical re-runs from a local store without consuming credits (8 of 20). Parse needs only a document. Official Python and TypeScript libraries and a CLI with JSON output (15).

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.

## 4. Transparency & trust, 63 out of 100, up to 3.2 more on the total

Made of editorial 49, provenance 76.

Why it scored 63: Closed service under published terms, with Apache-2.0 libraries and CLI (15 of 30). The Zero Data Retention page gives deletion timing per call type and names the regions. Statements disagree elsewhere. The terms allow training on customer materials with no exception written in them, while the Zero Data Retention page says no training when it is on. Retention without it is not stated, and the privacy policy gives a general seven-year period (14 of 30). DPT-1's retirement was announced in the changelog with a date of 31 March 2026, and models are pinned by dated snapshot. No written deprecation policy was found, and the general terms of service allow changes without notice (10 of 20). Hosting regions are named (AWS us-east-2 and eu-west-1). A sub-processor page exists in the trust centre, which our reader could not read. The CLI's usage telemetry is documented in its repository with an opt-out (10 of 20).

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: landing.ai, registered 2017-12-15 (8 years) (11 of 15)
- Terms of service: read, states 5 of the 7 things a reader expects, and has 3 clauses that cost points (2.3 of 10)
- Privacy policy: read, states 5 of the 8 things a reader expects (7.8 of 10)
- security.txt: not found (0 of 10)

## 5. Reliability, 85 out of 100, up to 3 more on the total

Why it scored 85: Graded as a hosted API. Instatus page at status.landing.ai with US and EU components for the product and an incident history feed (20). The feed shows no incident in the last 90 days. The most recent is a parsing outage of 18 hours 55 minutes on 30 April 2026, attributed to a cloud provider (30). Page limits per plan and tier are published, 100 pages a minute and 6,000 an hour on Explore (15). The docs ask for exponential backoff on 429, 502 and 503, say which 429 case will not clear on retry, and the libraries retry twice by default. No `Retry-After` header and no idempotency key for job creation were found (10 of 15). SLA and uptime guarantees are listed for Enterprise only and no SLA document is published (0). The Gen2 API has been generally available since 24 July 2026 (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.

## 6. Schema & documentation, 88 out of 100, up to 2 more on the total

Why it scored 88: Public OpenAPI 3.1 description of the nine Gen2 operations, and a second for Gen1 (25). `llms.txt`, `llms-full.txt` and a Markdown twin of every page (10). The overview says when to use Parse, Extract and Ground, and a separate guide compares synchronous calls with jobs on each tier. Operation descriptions in the description file are short, and one list operation has none (15 of 20). Parameters carry defaults, enums and ranges, but the job status filter is a plain string, the description file declares no security scheme, and parse options travel as a form field (10 of 15). Code samples in cURL, Python and TypeScript, and troubleshooting pages for every status code and 422 message. The description file itself lists only 200, 202, 206, 404 and 422 (13 of 15). Versioned `/v2` path, dated model snapshots and a dated changelog with an RSS feed (15). We read the description file and Markdown twins, not the rendered reference.

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. Maintenance & community, 85 out of 100, up to 1.3 more on the total

Why it scored 85: The API changelog's latest entry is 22 September 2026, and the Python and TypeScript libraries were tagged on 17 September (30). Eight dated changelog entries since 11 July 2026 (20). A public changelog with RSS, a support address answered Monday to Friday, and a Discord server. Replies on GitHub issues were not sampled (10 of 15). Current official libraries in Python (1.18.1) and TypeScript (2.13.1) (15). Both library repositories run CI, production end-to-end tests and a workflow that syncs them with the API description (10).

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.

## 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 trust centre at trust.landing.ai is drawn by script, so the sub-processor list, compliance reports and any DPA were not read.
- unchecked: `/legal/msa-ade-US-2026-02-11` and `/legal/msa-ade-EU-2026-02-11` are disallowed by landing.ai/robots.txt. Whether an agreement of that date replaces the linked terms for self-serve customers is not established.
- unchecked: GitHub stars, npm and PyPI download counts, and replies on GitHub issues.
- unchecked: the documentation MCP server's tool definitions. It serves documentation and skills, not the product, so it is not graded.
- The terms at landing.ai/va-tos forbid accessing the Solution through any agent, robot or tool LandingAI does not supply, and forbid benchmarking. Recorded as a fact with no deduction. It matters before any probe is run.
- Default retention and training use for Explore accounts without Zero Data Retention are not stated in the docs. The terms permit training.
- The Gen2 overview says password-protected files are not supported, while Python library 1.18.0 added a password option for v2 parse. Which is current was not established.
- One request went to a mistyped docs address (`/dpt3/ade-members.md`), which answered 307 and was not followed. The members page was not read.
- The lead was right about the interface. It did not mention the CLI, the EU deployment or the split between Gen1 and Gen2 APIs.

## Weaknesses

- The Explore plan has one API key that cannot be deleted or revoked. A leaked key is replaced by emailing support
- The published terms let LandingAI use customer materials and output to train its models. Zero Data Retention, which stops that, is a Team and Enterprise setting
- The terms forbid accessing the Solution through any agent, robot or tool LandingAI does not supply, and forbid benchmarking. This matters before any probe is run
- No idempotency key on job creation and no `Retry-After` header in the reviewed documentation
- No scoped or read-only keys, and no prompt-injection guidance for parsed document text was found
- The 18-hour-55-minute parsing incident of 30 April 2026 and two in February and March 2026 were attributed to a cloud provider

## 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.

- Send `Authorization: Bearer <key>` to `https://api.ade.landing.ai/v2/parse`. EU keys work only on `https://api.ade.eu-west-1.landing.ai`
- Submit large files to `/v2/parse/jobs`. Synchronous parse times out at 475 seconds and is capped at the plan's pages-per-minute limit
- Leave jobs on the `standard` tier unless speed matters. It consumes half the credits of `priority`
- Run Extract on the Markdown that Parse returned, then call Ground for page numbers and boxes. Ground consumes no credits
- On a 429, read the message. One case clears with backoff and the other means the document is too large for the tier

## 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.