# Fix list: Kontent.ai From Anchor Terminal's listing at https://www.anchorterminal.com/tools/kontent-ai, the October 2026 research run, assessed 9 October 2026. Grade BB, 70.5 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 Kontent.ai: 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. Payments & pricing, 20 out of 100, up to 10 more on the total Why it scored 20: Read with the hosted rubric. No x402, MPP or L402 (0). The pricing page publishes no figure. It shows sliders and a form that requests a calculation, and the terms place prices in an order form (0). The 30-day trial needs no card and includes API access. We did not run the sign-up form (20). A person signs up in a browser, enables the Management API and creates the key in 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. ## 2. Agent ergonomics, 71 out of 100, up to 4.7 more on the total Why it scored 71: Scored on the API. No field selection on Management API reads was found, and the docs send filtering to the Delivery APIs. A filter endpoint returns references only and a bulk-get endpoint fetches the chosen items. The MCP server loads 54 tools (13 of 25). Lists page by continuation token, and items with variants can be filtered. No page-size control was found (14 of 20). Errors carry a request ID, a code, a message and validation errors with path, line and position (17 of 20). No idempotency keys. Items can be addressed and upserted by external ID or codename, the docs say failed requests are safe to retry, and they advise against parallel requests. All 54 MCP tools set readOnlyHint and destructiveHint (13 of 20). Two official SDKs, JavaScript and .NET, and a call needs only the environment ID and a key (14 of 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. ## 3. Schema & documentation, 74 out of 100, up to 4.2 more on the total Why it scored 74: No OpenAPI file was found on the docs pages read. The reference links a Postman collection in a public workspace, which we did not read, and the MCP server's tools have typed schemas in its source (15 of 25). kontent.ai/llms.txt exists and is an index of marketing and docs pages. No Markdown copies of the docs were found, only PDF downloads (5 of 10). Each endpoint states its purpose and its limits, for example that scheduling a published variant has no effect and which endpoint to use instead (15 of 20). Attributes are typed with required, read-only and write-only flags, formats and length limits. Element values vary by element type (12 of 15). Each operation has request samples in TypeScript, C# and cURL, response status codes and an error object with validation paths (13 of 15). The version is in the path, product updates are dated with an RSS feed, breaking changes can be had by email, and a page lists which changes are not treated as breaking (14 of 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. ## 4. Reliability, 82 out of 100, up to 3.6 more on the total Why it scored 82: Read with the hosted lines and scored on the Management API. status.kontent.ai is a Statuspage site with components by region, the Management REST API among them, and an incident history (20). Its feed lists three incidents between 11 July and 9 October 2026. Logins to the web app were degraded for about an hour on 29 July with the APIs unaffected, the GraphQL service was disrupted for about 31 minutes on 16 September with the Management API unaffected, and content item saves failed for about 40 minutes on 24 September for some EU projects. We read those as minor (20). Rate limits have numbers, 10 requests a second and 400 a minute per environment (15). A 429 carries Retry-After, the docs call each failed request safe to retry and ask for a retry policy on 5xx. There are no idempotency keys, though the upsert PUT by external ID repeats safely (13 of 15). Annex 4 of the terms defines availability with a refund formula, but the guaranteed level is in each order form and the definition names the administration interface and the Delivery API, not the Management API (4 of 10). Management API v2 is generally available. The MCP server is at 0.39.0 (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. ## 5. Security & auth, 81 out of 100, up to 3.3 more on the total Why it scored 81: Management API keys have selectable permissions, environment limits, an expiry from 1 minute to 2 years (6 months by default), email reminders before expiry and revocation by regenerating. Keys travel only in the `Authorization` header. There is no OAuth (27 of 30). A key can be read-only, the API follows the workflow transition rules, and the MCP server marks 22 tools destructive so a client can prompt. Neither surface has its own confirmation step (15 of 20). Content is written by a project's own users and can carry instructions. The MCP README has a section on indirect prompt injection that recommends a least-privilege key and a human in the loop. The API docs say nothing on it (10 of 15). The audit log keeps 90 days and names the API key behind each change, with content item changes included since 2 October 2026. No API for reading it was found (12 of 15). security.txt is valid to 21 July 2027, the disclosure policy has a safe harbour and discretionary rewards, a penetration testing policy lets customers test, and the trust centre lists SOC 2 Type 2 and ISO 27001, 27017, 27018 and 42001. Reports are gated and no public advisories were found. We did not search NVD (17 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. ## 6. Transparency & trust, 75 out of 100, up to 2.2 more on the total Made of editorial 69, provenance 80. Why it scored 75: The service is closed under Terms of Service effective 11 September 2026. The SDKs and MCP server are MIT (15). The terms include a Data Processing Addendum and an AI annex that rules out training on customer content, and the privacy policy of 2 June 2025 covers product users. The periods differ by document. Content is kept 30 days after a subscription ends, the addendum says personal data is returned or deleted within 14 days, and the privacy policy keeps account data up to 90 days (23 of 30). No deprecation policy with a notice period was found. Removals are announced with dates, such as Sitemap on 14 July for 30 September 2026, and the terms give 14 days to object to a material change (13 of 20). The sub-processor page names each vendor with its address, location and the data sent, Anthropic and Microsoft among them, and the four data centre regions are documented (18 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: kontent.ai, no registry record we could read (0 of 15) - Terms of service: read, states 7 of the 7 things a reader expects, and has 2 clauses that cost points (6 of 10) - Privacy policy: read, states 7 of the 8 things a reader expects (9.3 of 10) ## 7. Maintenance & community, 83 out of 100, up to 1.5 more on the total Why it scored 83: The newest product update is dated 2 October 2026, the JavaScript SDK 8.5.7 was tagged on 21 September and the MCP repository's last commit is 5 October (30). Nine product updates fall between 14 July and 2 October 2026, and the MCP server had four tagged releases from 24 July to 31 August (20). Closed service with a dated changelog and RSS feed, in-app chat stated as 24/7 on all plans and a Discord community. We did not test a support channel (11 of 15). Official SDKs for JavaScript and .NET with releases since July 2026, and the MCP server's server.json names io.github.kontent-ai/mcp-server. We did not query the MCP registry (15). The repositories have build and test workflows and pinned, recent dependencies. 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. ## 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 Postman collection the API reference links. We read the rendered reference pages and not the collection - unchecked: the product updates page, which is drawn by script. We read the RSS feed it links - unchecked: GitHub star counts, open issues and CI results for the MCP server and SDK repositories. We read clones only - unchecked: whether io.github.kontent-ai/mcp-server is live in the official MCP registry. Only the repository's server.json was read - unchecked: the sign-up form itself. The trial page says no credit card is required - unchecked: NVD and GitHub advisories for Kontent.ai packages - unchecked: the domain registration date for kontent.ai - unchecked: the SOC 2 report, penetration test report and other trust centre documents, which need an access request - The Terms of Service forbid accessing the Service through automated processes unless the procedure is described in the documentation, and forbid API use other than as documented. Recorded as a fact with no deduction. It matters before any probe is run - The terms name a Developer Plan alongside the trial. The pricing page shows no such plan, and whether it can still be had was not established - The limits page says Management API requests count towards the plan's API call limit and also says there is no limit on the total number of API calls. The plan numbers are not public - Whether the Management API falls under the availability commitment was not established. Annex 4 names the administration interface and the Delivery API - One address we tried, https://kontent.ai/learn/llms.txt, does not exist. Only https://kontent.ai/llms.txt does - The privacy policy is issued by Kontent s.r.o. and the terms by Kontent CZ s.r.o., both at the same Brno address ## Weaknesses - No price is published. The pricing page shows sliders for seats, content types and items and a form that requests a quote - No OpenAPI file was found. The docs link a Postman collection, which we did not read - Access needs a person. The 30-day trial needs no card, but sign-up, the Management API toggle and key creation happen in the web app - No idempotency keys. The docs advise against parallel requests, saying they may cause inconsistencies in content - The availability annex covers the administration interface and the Delivery API, not the Management API, and the guaranteed level sits in each order form - The Terms of Service forbid automated access to the Service unless the procedure is described in the documentation. This matters before any probe is run - The MCP server loads 54 tools, 22 of them marked destructive, and runs only on the owner's machine. No hosted endpoint was found ## 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 ` to `https://manage.kontent.ai/v2/projects/{environment_id}/...`. The Management API must be switched on under Environment settings, General, Enabled APIs first - Ask for a Management API key with only the permissions the task needs. Read content alone gives a read-only key, and keys expire after six months by default - Stay under 10 requests a second and 400 a minute per environment, send requests one at a time, and wait the seconds in Retry-After on a 429 - Page lists by passing the `continuation_token` from the response back in the `x-continuation` header - Create items with an `external_id` and write with the upsert PUT at `/items/external-id/{id}` so a retry after a lost response does not create a duplicate - With the MCP server, call `get-patch-guide` before any patch tool, and use a client that prompts on destructiveHint ## 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.