Diffbot

# Fix list: Diffbot

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/diffbot, the October 2026 research run, assessed 9 October 2026. Grade C, 54.3 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 Diffbot: 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, 21 out of 100, up to 13.8 more on the total

Why it scored 21: One token per account, sent in the URL query string as the documented and only scheme in the OpenAPI file. Tokens are revocable in the dashboard per a June 2025 changelog title, and child tokens exist on Plus and above (20 of 30, less 10 for the secret in the URL, so 10). Child tokens have the same rights as the primary one and there is no read-only mode. The skills limit the agent to the `db` binary through `allowed-tools` (5 of 20). The API returns untrusted web content and no guidance on prompt injection was found (0 of 15). The token details page shows usage activity and an Account API reports usage. No per-call log was found (6 of 15). No security.txt, security page, certification or bug bounty was found on the pages read, and the privacy policy says some API calls rely on HTTP without encryption (0 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. Reliability, 55 out of 100, up to 9 more on the total

Why it scored 55: Graded as a hosted API. status.diffbot.com is a Pingdom public report with 13 checks, among them Analyze API, Article API, Crawlbot and Websearch, with daily uptime for seven days and no incident notes (15 of 20). No 90-day record can be read. In the week to 9 October 2026 Analyze API shows 99.69 per cent, with 2, 11 and 17 minutes down on 4, 5 and 6 October, and Article API 5, 9 and 2 minutes on 5 to 7 October (5 of 30, the line for no readable history). Rate limits are published per plan, 5 calls a minute on Free, 5 a second on Startup and 25 a second on Plus (15). The 429 page says to pause one second and back off. No `Retry-After` header is documented, though the official client reads one (10 of 15). No SLA was found (0). The Extract APIs are at `/v3` with no beta label (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.

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

Why it scored 40: No x402, MPP or L402 was found (0 of 40). Plan prices, the credit cost of each action and overage per credit are public without a login (20). The Free plan has 10,000 credits a month and needs no card (20). A person signs up in a browser at app.diffbot.com to get a token (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, 63 out of 100, up to 6 more on the total

Why it scored 63: Graded on the REST API, with the skills noted. `mode` limits extraction to one page type, `fields` adds optional fields and `discussion=false` drops comments, but an article response carries both `html` and `text` with no way found to drop either. The skills return Markdown by default (14 of 25). A single-URL call has no pagination. Web Search takes `size` and `maxTokens` and DQL takes `size` per the client. Crawl result paging was not read (10 of 20). Errors are JSON with `errorCode` and `error` and six docs pages. A failed extraction can arrive in a 200, and the 429 text says the site received too many requests when the cause is the caller's plan limit (12 of 20). Extract calls are GET reads and safe to retry. Whether a failed call uses a credit was not found, and there is no idempotency key for creating crawl jobs or custom APIs (12 of 20). Only `url` and `token` are required, with official Python and TypeScript clients and a CLI (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.

## 5. Maintenance & community, 57 out of 100, up to 3.8 more on the total

Why it scored 57: The official Python client was tagged v3.0.0 on 28 September 2026, eleven days before the check. The API changelog's newest entry is 15 June 2026, 116 days before the check, and recency is scored on that entry (10). Tagged releases in the last 90 days are skills v1.1.0 and v1.1.1 on 12 and 13 August and Python v3.0.0 on 28 September. The API changelog has none in that window (20). A public changelog with RSS and a support address exist. GitHub issues were not read (8 of 15 for a closed service). Python and TypeScript clients are current, but the docs name the packages `diffbot-python` and `diffbot-typescript` while the repositories publish `diffbot` and `@diffbot/typescript` (12 of 15). The Python repository tests on Python 3.9 to 3.13 and raised two dependencies for security fixes on 28 September. We did not see CI results (7 of 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.

## 6. Schema & documentation, 83 out of 100, up to 2.8 more on the total

Why it scored 83: The Extract APIs have a public OpenAPI 3.1.0 file, version 1.1.0, with 13 operations, and the reference pages are generated from it. We read that file and one rendered page. The descriptions for Crawl, Knowledge Graph and Web Search were not read (25). llms.txt exists but is a short index with no Markdown twins of the docs pages found (5 of 10). The Analyze API page has sections on when to use it and when not to (17 of 20). Parameters are typed, `url` is the one required field and the file has 15 enums, but `fields` is a comma-separated string and the file defines no shared response schemas (10 of 15). The file carries 24 examples. It documents only 200 and 500, with separate docs pages for 401, 404, 429 and 457 (11 of 15). The path is versioned `/v3` and a dated changelog with RSS runs from 2015 to 15 June 2026 (15).

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, 68 out of 100, up to 2.8 more on the total

Made of editorial 49, provenance 86.

Why it scored 68: A closed service under undated Terms of Use that cover the Site and the Service, cap liability at $10 and grant Diffbot a perpetual licence to process submitted queries. The clients and skills are MIT (13 of 30). The privacy policy of 29 August 2025 says query history and API calls are stored against the API key and kept no longer than necessary, with no retention period given. The GDPR and data rights page was not read (14 of 30). There is no deprecation policy. The changelog carries dated notices, such as a May 2024 entry announcing a field removal for 30 August 2024, and the terms give 14 days' notice of fee changes (8 of 20). The privacy policy names its service providers one by one, among them AWS, Stripe, Intercom, Salesforce, Slack and Amplitude, and says all are in the United States (14 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):

- Terms of service: read, states 5 of the 7 things a reader expects, and has 1 clause that costs points (6.3 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 docs pages for Crawl, Bulk Extract, Web Search, DQL, Enhance, Natural Language, credits, custom JavaScript, proxies and the remaining error pages. We kept to about twenty requests on www.diffbot.com, the standard files and the OpenAPI file included, which is over the brief's fifteen
- unchecked: the GDPR and data rights page, the cookie policy, the EEA and California privacy pages and the support page, so whether a DPA is published was not established
- unchecked: a 90-day incident record. The Pingdom report shows seven days
- unchecked: GitHub stars and issue responsiveness. api.github.com answered 403 for a rate limit
- unchecked: whether `diffbot-typescript`, the name in the docs, exists on npm. The repository's package is `@diffbot/typescript`, and no release tag was found in it
- Whether a failed or timed-out extraction uses a credit
- Whether JavaScript rendering and residential proxies are available. The pages read document a data centre proxy and a caller's own proxy only
- The terms forbid software that scrapes, spiders or crawls the Site or Service. That matters before any probe is run
- The lead was right on the interface. It gave the vendor as Diffbot. The legal entity is Diffbot Technologies Corp.

## Weaknesses

- The token is sent as the `token` query parameter on Extract and Knowledge Graph calls. It is the only scheme in the OpenAPI file
- Child tokens have the same access rights as the primary token, and creating them needs the Plus plan ($899 a month) or a request to support
- No security page, certification, bug bounty or security.txt was found on the pages read. The privacy policy says some API calls rely on HTTP without encryption
- The status page is a Pingdom report with seven days of uptime per check and no incident notes. Analyze API shows 99.69 per cent for the week to 9 October 2026
- The docs give `pip install diffbot-python` and `pnpm add diffbot-typescript`, while the repositories now name the packages `diffbot` and `@diffbot/typescript`
- The terms grant Diffbot a perpetual, irrevocable licence to process submitted queries, cap liability at $10, and forbid software that scrapes or crawls the Site or Service. That clause matters before any probe is run

## 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.

- Call `GET https://api.diffbot.com/v3/analyze?url=<encoded URL>&token=<token>` when the page type is unknown. Use `/v3/article` or `/v3/product` to force one schema
- Keep the token out of logs and shared URLs. It sits in the query string, so any proxy or log that records the request URL records the token
- Stay under 5 calls a minute on Free, 5 a second on Startup and 25 a second on Plus. On 429 wait at least one second and back off
- Check a 200 response for `errorCode`. The official client raises `ExtractionError` when a 200 reports a failed extraction
- Set `useProxy=default` only when a site blocks the fetch. A proxied page costs 2 credits against 1
- Install `diffbot` from PyPI, not `diffbot-python`, which stopped at 0.3.0

## 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.