Financial Cents MCP

# Fix list: Financial Cents MCP

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/financial-cents-mcp, the October 2026 research run, assessed 27 October 2026. Grade E, 39.7 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 Financial Cents 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, 23 out of 100, up to 15.4 more on the total

Why it scored 23: Graded with the hosted lines, on the remote MCP server. No status page was found (status.financial-cents.com doesn't resolve) (0 of 20), so there's no readable incident history (5 of 30). Financial Cents says MCP requests are rate limited but gives no number; the REST API's 250 requests a minute is for a separate surface (0 of 15). No 429 or Retry-After guidance, but the tool descriptions give safe-retry rules for writes. Don't retry a successful send or charge, send `expected_amount_remaining` so a credit-memo retry can't apply twice, and upload IDs are single use (8 of 15). No SLA found in the pages we could read; the terms weren't readable (0 of 10). The changelog calls it an initial release and nothing labels it beta (10 of 10). Total 23.

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. Schema & documentation, 40 out of 100, up to 9.8 more on the total

Why it scored 40: No machine-readable contract could be read. `tools/list` answers 401 without a signed-in Scale account and no public manifest or schema exists (0 of 25). No llms.txt on either host and no Markdown docs for the MCP; the gitbook llms.txt covers the REST API only (0 of 10). Every tool's full description is public on the setup guide, and they say what each tool is for, which tool to use instead, what it won't do and the units of money fields (18 of 20). The descriptions name enums such as `resource`, `board`, `group_by` and `log_type` and the required fields on create, but the typed schemas weren't read (6 of 15). Example prompts mapped to tools, a troubleshooting section, and descriptions that say which inputs are rejected and why; no error shape (8 of 15). A public changelog with a dated entry for the initial release, but no server version (8 of 15). Total 40.

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.

## 3. Payments & pricing, 25 out of 100, up to 9.4 more on the total

Why it scored 25: No x402, MPP or L402 (0 of 40). Scale plan prices are public, $69 a user a month annually or $89 monthly, but the MCP has no price of its own and no row on the pricing page, so plan-only pricing (10 of 20). A 14-day trial with no card. Judgement call, 15 of 20, because the release points new firms to the trial but no page says the MCP is on during it, and the MCP page says support may need to enable it (15 of 20). A person has to sign up and sign in through a browser (0 of 20). Total 25.

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, 48 out of 100, up to 8.5 more on the total

Why it scored 48: Graded on the MCP server. 40 tools, more than 30, with about 30,000 characters of descriptions on the guide, the longest about 2,100, before schemas. No toolsets or read-only subset (5 of 25). List tools take filters, `limit`, pagination, `include` for nested data and `group_by`, with totals over the whole filtered set (18 of 20). Descriptions say what gets rejected and that the reason is returned, such as a locked time entry or a proposal reminder sent too soon with the next allowed time; no error codes documented (10 of 20). No idempotency keys, but `preview` on proposal sends, `expected_amount_remaining` on credit memos and single-use upload IDs. The guide labels tools Read or Write, but readOnlyHint and destructiveHint couldn't be checked (8 of 20). Omitted fields are left unchanged on updates and `whoami` grounds dates; no official SDK (7 of 15). Total 48.

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. Security & auth, 61 out of 100, up to 6.8 more on the total

Why it scored 61: OAuth 2.1 with PKCE S256 and dynamic client registration, tokens bound to one firm and revocable in the app, with no API key. One scope, `mcp:use`, covers every read and write, so 22 rather than 30 (22 of 30). No read-only mode. The user's app permissions apply and no tool deletes, but the server can send, charge and void invoices and email clients. Server-side flags exist only for uploads that email portal contacts (`confirm_notify`) and recurring back-charges (`confirm_back_charges`); other confirmations are left to the AI client, guided by the descriptions (10 of 20). Client-written chat and topic names come back inside `<untrusted>` tags, with injection guidance on both pages (13 of 15). Financial Cents says every call is logged with who, what and how long, and connected clients are listed in Settings; no customer view of the call log is described (8 of 15). No security.txt (404 on both hosts) and no bug bounty found. The security page claims SOC II Type I and twice-yearly third-party assessments, with an attestation letter on request (8 of 20). Total 61.

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. Maintenance & community, 50 out of 100, up to 4.4 more on the total

Why it scored 50: Closed hosted service. The changelog dates the initial release to 5 October 2026, 22 days ago (30 of 30). One dated MCP entry in 90 days; the tool count grew from 33 in the recap of 19 October to 40 now, without dated notes (10 of 20). A public changelog, roadmap and feedback board on Canny, an Intercom help centre with an MCP article updated 23 October, and a Facebook community (10 of 15). Not in the official MCP registry (searches for financial-cents and financialcents returned nothing on 27 October 2026) and no SDK (0 of 15). No package (0 of 10). The server is about three weeks old. Total 50.

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.

## 7. Transparency & trust, 53 out of 100, up to 4.1 more on the total

Made of editorial 30, provenance 76.

Why it scored 53: The editorial half. A closed service whose terms sit on iubenda behind a robots.txt disallow, so they weren't read (10 of 30). The privacy page shows a 2018 website policy and links an iubenda policy that wasn't readable. The security page says deleted records are removed within six months and vault records at once. Nothing was found on how MCP call logs or assistant data are kept (10 of 30). No deprecation policy or dated notices; the REST API docs warn breaking changes could happen (0 of 20). The security page names DigitalOcean and AWS with servers in New York City and Virginia, and the privacy page names Google Analytics; no full subprocessor list (10 of 20). Total 30.

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: financial-cents.com, registered 2017-04-10 (9 years) (11 of 15)
- Status page: not found (0 of 10)
- security.txt: not found (0 of 10)

## Deductions

Each comes off the total. A fixed and documented problem counts for less at the next check.

- -2, misleading claim. The MCP page and its FAQ (last modified 15 October 2026) say the server can view invoices, payments and proposals, while the setup guide lists `manage_invoice`, `invoice_actions` (send, finalise, charge, void) and `proposal_actions` (send). That understates the server's reach over client money. Checked 27 October 2026 (https://financial-cents.com/mcp/, https://app.financial-cents.com/mcp-guide).

## 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 tools as served, because `initialize` and `tools/list` answered 401 without a signed-in Scale account. The count of 40, the names and the descriptions are from the setup guide, which says they're read from the code; input schemas and annotations are unread.
- unchecked: the terms of use and the current privacy policy, both iubenda documents that iubenda's robots.txt disallows. The privacy page shows a 2018 website policy.
- unchecked: whether the 14-day trial includes the MCP. The release points new firms to the trial, the MCP page says support may need to enable MCP for a firm, and the pricing page has no MCP row.
- firstReleased is 5 October 2026, the date of the changelog entry for the initial release, which says it shipped "this past week". The September recap also places the launch in September. Both are earlier than the release of 20 October, which is the lead's date. The MCP page was last modified on 15 October per the sitemap.
- The September recap says Claude users can search for Financial Cents in their connectors, but no Financial Cents entry was found in claude.com's connector sitemap, and the setup guide uses a custom connector. Recorded, not deducted.
- The deduction for the MCP page's view-only description of invoices, payments and proposals may lapse if the page is updated to match the guide; confirm at landing.
- Category. project-management, because tasks, projects, time and client work are the server's main jobs. About a third of the tools are billing (invoices, payments, proposals, credit memos, tax codes), so accounting was the alternative, and crm.records and accounting.invoices are listed as capabilities. Judgement call.
- Our own products. Financial Cents MCP doesn't do LocalGhost's job or letme's. It reaches one vendor's practice-management data and picks no tools. Graded the normal way.

## Weaknesses

- One OAuth scope, `mcp:use`, covers reads and every write, including sending, charging and voiding invoices
- The MCP marketing page says invoices, payments and proposals are view-only, while the guide lists tools that send, charge and void them
- No status page, SLA or numeric rate limit for the MCP; the REST API's 250 requests a minute is a separate surface
- Input schemas and annotations aren't public, since `initialize` and `tools/list` need a signed-in Scale account
- Scale plan only, at $69 a user a month billed annually; no free plan and no MCP row on the pricing page

## 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 `whoami` first and work out date windows from `user.timezone`, not `company.timezone`
- Resolve clients and projects with `list_clients` or `list_projects` search, since `lookup` covers only reference data such as users, tags and service items
- Check money units per tool. Invoices, payments, proposals and credit memos use integer cents, while time entries and service-item rates use dollars
- Never retry a successful `invoice_actions`, `proposal_actions` or credit-memo create, since a second call can email or charge the client again
- Get the person's approval before any write, and use `preview: true` before sending a proposal

## 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.