# Fix list: Paychex From Anchor Terminal's listing at https://www.anchorterminal.com/tools/paychex, the October 2026 research run, assessed 8 October 2026. Grade E, 44.8 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 Paychex: 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, 0 out of 100, up to 12.5 more on the total Why it scored 0: Read with the hosted rubric. No x402, MPP or L402 (0). No price is published for the API or for payroll. The Request Pricing button leads to the sales contact form (0). No free tier or trial was found. The sandbox is issued only after Paychex approves a partnership, and a client needs a Paychex Flex account (0). A person fills in a questionnaire or creates the app inside Flex, and Paychex issues the credentials (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. Reliability, 41 out of 100, up to 11.8 more on the total Why it scored 41: Read with the hosted lines and scored on the Paychex API at https://api.paychex.com. No status page was found on the developer centre or the security page, and status.paychex.com didn't resolve (0). With no page there is no readable incident history (5). The rate limiting page gives 10,000 requests a minute per partner key and a burst of 2,000 a second. It is headed coming soon for 25 September with no year, and we couldn't confirm the limit is live (15). 429 responses carry `Retry-After` and three `X-RateLimit` headers, and 423 tells a caller the client is locked and to retry later. The `X-payx-client-correlationId` header is said to be usable for idempotency, with no rules for how a repeated ID is treated (11 of 15). No SLA was found, and the User Terms of Use say Paychex doesn't guarantee availability (0). The API is version 1.0 with no beta or preview 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. Agent ergonomics, 54 out of 100, up to 7.5 more on the total Why it scored 54: Scored as an HTTP API. Lists are sized with `offset` and `limit` on three endpoints, and vendor media type profiles select a response shape, among them non-PII worker profiles. No field selection (13 of 25). Paging exists on companies, company checks and company workers only. Workers can be filtered by name, employee ID, location, status and creation dates, but the reference says paging and filters can't be combined (11 of 20). Errors carry `code`, `description` and `resolution` and are listed per operation (16 of 20). A correlation ID header, `workerCorrelationId` on batches, 207 for partial success and 423 on a locked client help retries. No idempotency key with stated behaviour was found, and there are no MCP annotations to read (8 of 20). Each resource has a default media type and most calls need only a path ID. No SDK was found in any language (6 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. ## 4. Security & auth, 58 out of 100, up to 7.4 more on the total Why it scored 58: OAuth 2.0 client credentials with the key and secret in the request body, 60-minute Bearer tokens and no refresh token. A client chooses an app's access levels in Flex, a partner app is limited to the verbs and endpoints Paychex grants, and the sample token carries scopes such as `read:company_people`. No way to rotate a secret was found in the reviewed pages (24 of 30). We took 5 off, not the checklist's 10, because the webhook option `OAUTH2` sends the receiver's own client ID and secret as query parameters and the API's credential never travels in a URL (19). A client admin approves each partner app in Flex and access can be removed. Checks are created unprocessed and a person submits the payroll. Direct deposit and tax changes apply without a second step (16 of 20). Responses include free text such as memos, custom fields and documents, and no guidance on untrusted content was found (3 of 15). Every response has an `x-payx-txid`. No per-call log for the operator was found (5 of 15). A responsible disclosure policy with a form run with Bugcrowd and no monetary award, SOC 1 Type 2 and SOC 2 Type 2 reports on request, ISO 27001 certificate IS 801702. No security.txt (15 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. ## 5. Maintenance & community, 26 out of 100, up to 6.5 more on the total Why it scored 26: The OpenAPI file was last modified on 20 August 2026, 49 days before this check. The two News notices give no year for their own dates (20 of 30). Three dated changes in the last 90 days couldn't be established (0). Closed service with a support form split into five request types, an FAQ and a chat widget on the developer centre. No public forum or issue tracker was found (6 of 15). No official SDK was found (0). No packages or public CI to assess (0). 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, 66 out of 100, up to 5.5 more on the total Why it scored 66: One public OpenAPI 3.0.3 file with 68 paths, 110 operations and 147 schemas, drawn by Redoc on the documentation page (25). developer.paychex.com/llms.txt returns 404 and the guides are HTML only. The llms.txt on www.paychex.com lists marketing pages (0). Operation descriptions have a median length of 63 characters. A few, such as adding in-progress workers, explain statuses and limits, and most restate the summary without saying when not to use the call (11 of 20). 48 enums, 244 required lists and 162 formats, with no patterns or length limits, and the file declares no security scheme or scopes (10 of 15). About 250 named request and response examples, and each operation lists error responses with codes such as `API-40` and `API-10` (13 of 15). Versioning by vendor media type is documented. The News page holds two notices and there is no dated changelog (7 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. ## 7. Transparency & trust, 54 out of 100, up to 4 more on the total Made of editorial 39, provenance 69. Why it scored 54: Closed service. The Paychex User Terms of Use, effective 1 June 2026, are public and cover Paychex's sites, applications and services. The partner agreement and the Third-Party Terms of Use a client accepts in Flex weren't found in public (12 of 30). The privacy policy, last updated 1 June 2026, covers data Paychex holds as a service provider for clients and says personal information isn't sold for third-party marketing. Retention is stated as long as necessary with no periods, the terms let Paychex use submitted information to improve its services and AI tools, and no public DPA was found (12 of 30). Deprecations are announced with dates. The unpaged worker list ends on 24 February 2027 and the versioning page promises never to drop fields within a version. No notice period is stated, and the reference still warns that POST federal tax will be deprecated on 30 January 2026 (12 of 20). No sub-processor list or data locations were found. The webhooks page names three production sites as WDC, HDC and ODC (3 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 3 clauses that cost points (2.3 of 10) - Privacy policy: read, states 6 of the 8 things a reader expects, and has 1 clause that costs points (6.5 of 10) - 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: whether rate limiting is live. The page is headed coming soon for 25 September with no year, and the unauthenticated 401 we received carried no `X-RateLimit` headers - unchecked: the partner agreement and the Third-Party Terms of Use a client accepts in Flex. Neither was found in public, so the terms link is the User Terms of Use - unchecked: any fee Paychex charges partners or clients for API access. No figure is published - unchecked: the 2026 overview PDF of Paychex's security programme linked from the security page, which we didn't read - unchecked: Paychex's GitHub organisation. github.com/paychex exists, and the GitHub API refused us for its rate limit, so whether any repository is an API client is unconfirmed. The developer centre names none - unchecked: the Paychex Marketplace page, which is drawn by script - Whether a status page exists under another address. None is linked from the developer centre, the security page or the support form - Whether `X-payx-client-correlationId` makes a repeated POST safe. The headers page says it can be used for idempotency and gives no rules - The security page lists bug bounty programmes among its assessments, while the responsible disclosure page says the programme pays no award - The User Terms of Use bar robots and other automatic means of access unless Paychex approves, and bar using the sites to develop or train AI. How this applies to an approved API app isn't stated. Recorded as a fact with no deduction - www.paychex.com/llms.txt carries usage guidelines addressed to AI models. We read it as data and didn't act on it - The listing omits `payroll.run` because the API enters checks but can't submit a payroll - The OpenAPI file's name includes develop.210 and its examples list pre-processing and AI option resources that have no public path ## Weaknesses - Partner access needs a questionnaire, review by the Corporate Partnerships Team, a demo and final approval. The sandbox is issued only after approval - No price is published for the API or for Paychex payroll. The pricing button leads to a sales contact form - No status page, incident history or SLA was found, and the User Terms of Use give no availability guarantee - Paging works on three endpoints only, and on the worker list it can't be combined with filters - No SDK, llms.txt or dated changelog was found. The News page holds two notices - The User Terms of Use bar automated access to Paychex sites and applications unless Paychex approves it, and allow suspension without notice ## 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. - POST `grant_type=client_credentials` with the key and secret in the form body to `/auth/oauth/v2/token`. Tokens last 60 minutes and there is no refresh token, so request a new one - Call GET `/companies` first. Worker and payroll paths need a `companyId` or `workerId` the app has been granted - Send `offset` and `limit` on GET `/companies/{companyId}/workers`, 50 at most, and resend the `ETag` between pages. The unpaged form is withdrawn on 24 February 2027 - Creating a check doesn't pay anyone. A person must review and submit the payroll in Paychex Flex - Send a unique `X-payx-client-correlationId` per request and keep the `x-payx-txid` from the response. Support asks for it - On 423 the client is locked by another user or process. Wait and retry ## 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.