Expensify

# Fix list: Expensify

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/expensify, the October 2026 research run, assessed 8 October 2026. Grade E, 41.1 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 Expensify: 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. Schema & documentation, 28 out of 100, up to 11.7 more on the total

Why it scored 28: No OpenAPI or other machine-readable spec was found. The reference is one HTML page plus a page for the Advanced Employee Updater (0). www.expensify.com/llms.txt and help.expensify.com/llms.txt both return 404 (0). Each of the 16 documented jobs has a one-line purpose and notes on the admin rights it needs, with little on when not to use it (10 of 20). Parameter tables give format, valid values and required or optional for each field. The request itself is a JSON document inside a form field, and exports need a Freemarker template (8 of 15). Every job has a curl example and a sample error, and three jobs have an error code table. Codes are unconventional, with 410 for validation errors and 404 for a failed sign-in (10 of 15). The API has no version number and no changelog was found (0).

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.

## 2. Security & auth, 34 out of 100, up to 11.6 more on the total

Why it scored 34: A partnerUserID and partnerUserSecret pair is generated per account at www.expensify.com/tools/integrations, shown once and replaceable there. It acts with all of the account's rights. OAuth2 for the API is in private beta with one scope, `integrations:api`, two-hour access tokens and rotating refresh tokens. The MCP server uses OAuth 2.1 with PKCE (18 of 30). A Postman guide on Expensify's help site places `requestJobDescription`, which carries the secret, in the URL parameters, so the checklist's 10 comes off (8). The MCP server is read-only. The API has no scopes or read-only credential, creating on behalf of other users needs Expensify to enable it for the domain, and no confirmation step was found for writes (8 of 20). Merchant names and comments are untrusted text, and no prompt-injection guidance was found (3 of 15). An `onFinish` email can report what a job changed. No log of API calls was found in the docs (3 of 15). Expensify states annual SOC 1 Type 2 and SOC 2 Type 2 audits, PCI DSS compliance and yearly penetration tests, with reports on request at trust.expensify.com. security.txt returns 404 and no disclosure policy or bounty was found (12 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.

## 3. Agent ergonomics, 45 out of 100, up to 8.9 more on the total

Why it scored 45: Exports return only the fields the caller's Freemarker template names, `limit` caps the number of reports, and the policy getter takes a `fields` list. An export is two calls, one to generate a file and one to download it (16 of 25). No pagination is documented. Filters cover report ID, workspace, start and end date, approved-after date, report state and an exported label (10 of 20). Errors are a `responseCode` and a `responseMessage` in the body, with specific messages such as a malformed date. There is no list of messages, and an invalid credential returned HTTP 200 with `responseCode` 404 when we tried it (9 of 20). No idempotency key is documented. The `markedAsExported` filter stops an export repeating, the report status updater returns 207 with lists of skipped and failed reports, and the employee updater has `dry-run` (5 of 20). One endpoint and few required fields per job. No official SDK was found (5 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. Payments & pricing, 30 out of 100, up to 8.8 more on the total

Why it scored 30: Read with the hosted rubric. No x402, MPP or L402 was found in the API docs or the pricing article (0). Prices are public per member, Collect at $5 a month and Control at $18 a month on an annual subscription or $36 pay-per-use. The API has no per-call price (10). A workspace starts a 30-day free trial and a payment card is asked for when it ends, and a free Submit workspace exists for individuals (20). A person has to create the account and open the integrations page in a browser to generate 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.

## 5. Reliability, 62 out of 100, up to 7.6 more on the total

Why it scored 62: Read with the hosted lines and scored on the Integration Server API, the surface an outside agent can write through. A Statuspage site at status.expensify.com lists 20 components, with Integration APIs as its own (20). In the 90 days to 8 October 2026 no incident names Integration APIs. The website and app had four incidents of over an hour, on 9 August (critical, 8 hours 45 minutes, described as degraded performance), 19 August (critical, 2 hours 44 minutes), 31 August (major, 1 hour 42 minutes) and 18 September (major, 1 hour 16 minutes). The updates don't say whether the API was affected, and OAuth sign-in runs on www.expensify.com, so we scored between minor only and one major (15 of 30). Limits are published, 5 requests per 10 seconds and 20 per 60 seconds (15). The docs name status 429 but give no Retry-After header or backoff guidance, and no idempotency key exists for create jobs. A `test` flag on exports and `dry-run` on the Advanced Employee Updater are the only safe-retry aids (4 of 15). No SLA was found in the Terms of Service (0). The API is generally available. OAuth2 partner authentication is in private beta and the older Employee updater is marked no longer maintained (8 of 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. Maintenance & community, 23 out of 100, up to 6.7 more on the total

Why it scored 23: No changelog or release record was found for the Integration Server API. The reference page carried a Last-Modified header of 8 October 2026, and its OAuth2 section is undated. The MCP help articles were first committed on 27 May 2026 and last revised on 28 September 2026. With no dated API change, half credit (15 of 30). No dated changelog entries (0). Support is Concierge by email or chat, and the help article says the API is self-serve and the customer's own team manages setup. Response times couldn't be observed (6 of 15). No official SDK was found, and the MCP server isn't in the official MCP registry (0). No package to assess. The API's response headers name GlassFish Server Open Source Edition 4.1.1 and Java 1.8 (2 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.

## 7. Transparency & trust, 58 out of 100, up to 3.7 more on the total

Made of editorial 41, provenance 75.

Why it scored 58: Closed service under the Terms of Service, last updated 10 September 2026. The Expensify app client is MIT on GitHub, and the API server is not public (15 of 30). The privacy policy (1 April 2026) says personal data is deleted as soon as practicable after an account ends, keeps data while an account is active, and lists California retention periods of up to 10 years by category. It states participation in the EU-U.S. Data Privacy Framework. No public data processing addendum was found on the pages read (17 of 30). The reference marks the Employee updater as no longer maintained with no end date, and the terms allow changes or discontinuation without prior notice (3 of 20). The terms and privacy policy name OpenAI, Spotnana and Aircall as providers. No full sub-processor list or data location statement was read, and the trust centre renders with JavaScript (6 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 6 of the 7 things a reader expects, and has 2 clauses that cost points (5.1 of 10)
- Changelog: 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: the pricing page at www.expensify.com/pricing renders with JavaScript. Prices come from Expensify's pricing help article
- unchecked: trust.expensify.com renders with JavaScript, so the sub-processor list, the security reports and any disclosure policy there weren't read
- unchecked: whether API credentials work on the free Submit workspace or during the trial without a paid plan. The reference only says to create an account
- unchecked: whether 429 responses carry a Retry-After header
- unchecked: the MCP server's tool list, input schemas and annotations, which need a signed-in account. The help article describes one Search tool
- unchecked: whether an SLA or a data processing addendum is available under a Statement of Work. Neither was found in the Terms of Service or the privacy policy
- unchecked: when the Integration Server API last changed. No changelog exists, and the Last-Modified header of 8 October 2026 may reflect a redeploy
- The lead named only the Integration Server API. Expensify also runs a read-only MCP server at https://www.expensify.com/mcp, and OAuth2 for the API is in private beta by request form
- The 10-point query string deduction rests on a customer-written Postman guide published on Expensify's help site. The reference's own examples send the secret in the POST body
- The four website incidents of over an hour in the last 90 days don't name the Integration APIs component, and their updates give no detail. The incident score of 15 is a judgement between 20 and 10

## Weaknesses

- No OpenAPI spec, llms.txt, version number or changelog was found for the Integration Server API
- The partnerUserID and partnerUserSecret pair acts with all of the account's rights. No scopes or read-only credential were found for the API
- No idempotency key or pagination is documented, and no official SDK was found
- Errors arrive as `responseCode` in the body. An invalid credential returned HTTP 200 with `responseCode` 404 when we tried it on 8 October 2026
- The Terms of Service allow Expensify to modify or discontinue the service at any time without prior notice, and no SLA 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.

- POST `requestJobDescription` as a form field to the single endpoint. Keep it out of the URL query string, because it carries the secret
- Read `responseCode` in the body on every reply. HTTP status can be 200 on failure, and validation errors use 410
- Stay under 5 requests per 10 seconds and 20 per 60 seconds. No Retry-After header is documented, so space retries yourself
- Export in two calls. A `file` job returns a filename, then a `download` job fetches it. Amounts are in cents
- Don't retry a create job blindly. No idempotency key exists, so export and check before resending. Set `dry-run` to true when testing the Advanced Employee Updater

## 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.