Fastmail API (JMAP)

# Fix list: Fastmail API (JMAP)

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/fastmail, the October 2026 research run, assessed 8 October 2026. Grade C, 54.1 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 Fastmail API (JMAP): 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, 54 out of 100, up to 9.2 more on the total

Why it scored 54: Graded on the JMAP API with the hosted lines. fastmailstatus.com, run on Instatus, lists ten components with notice history, though none is named for JMAP or the API (20). In the 90 days to 8 October 2026 the history shows three partial outages that stopped some users reaching the service, 8 hours 16 minutes from 25 August, 1 hour 8 minutes on 4 September and 5 hours 23 minutes on 7 October, plus 11 hours of bounced replies from masked addresses on 23 September. Each touched a subset of users, and we read the three together as one major outage (10). Sending limits are published per plan (8,000 messages a day on Individual and Standard, 120 on a trial, 500 logins per ten minutes), but no API request rate was found. JMAP servers state request limits in the session object, which we couldn't read without an account (8 of 15). RFC 8620 defines the `urn:ietf:params:jmap:error:limit` problem type and a `rateLimit` error, and the developer page says Masked Email creation returns it. No Retry-After or backoff guidance was found, and `EmailSubmission/set` has no idempotency key (6 of 15). No SLA was found in the terms of service, the API terms or the pricing page, and the API terms say the APIs come as is (0). JMAP access with an API token carries 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.

## 2. Payments & pricing, 30 out of 100, up to 8.8 more on the total

Why it scored 30: No x402, MPP or L402 on the developer page, the pricing page or the 401 responses from api.fastmail.com (0 of 40). Plan prices are public without a login, $6 a month or $60 a year for Individual and $6 a user a month for Business Standard, with API access included and no per-call price (10 of 20). A trial of up to 30 days needs no card, though Fastmail may ask for a mobile number, and the trial page doesn't say whether API tokens work during it (20). A person signs up in a browser and creates a token in settings, or consents on an OAuth screen (0 of 20).

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.

## 3. Schema & documentation, 52 out of 100, up to 7.8 more on the total

Why it scored 52: Graded on the JMAP API. Fastmail publishes no OpenAPI or similar file. The contract is the IETF text (RFC 8620, 8621 and 9610) plus a session object that lists capabilities and limits for a signed-in caller, which is machine-readable discovery but not a full schema. The MCP server's tool definitions sit behind sign-in and were not read (10 of 25). `www.fastmail.com/llms.txt` returned 404 on 8 October 2026 (0). The RFCs state the purpose, arguments, defaults and error cases of every method, and Fastmail's own page documents the scopes and the Masked Email extension. Nothing says when to prefer JMAP, IMAP or MCP (14 of 20). Arguments are typed in the RFCs with defaults and required fields, and filters are structured objects, with no schema file to validate against (12 of 15). The RFCs carry worked requests and error lists, and fastmail/JMAP-Samples has scripts in JavaScript, Python, Perl and Lua. Fastmail documents no errors of its own beyond `rateLimit` on Masked Email (11 of 15). Capability URIs version the protocol, but no API changelog was found. The API terms mention changelogs without a link, and `/changelog/` and `/whats-new/` returned 404 (5 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. Security & auth, 60 out of 100, up to 7 more on the total

Why it scored 60: Graded on the JMAP API with the MCP controls counted where a line covers them. OAuth 2.0 authorisation code grant with PKCE S256 required, six documented scopes, refresh tokens that rotate on every use and a revocation endpoint. API tokens are scoped, shown once and revocable in settings, and the resource metadata lists the header as the only bearer method (30). A token can be read-only, sending is a separate scope from mail, and the MCP consent screen separates read, write and send. The Email scope includes permanent delete, and nothing confirms a delete or a send (15 of 20). Mail is untrusted content. No guidance on prompt injection was found on the developer page, the MCP help article or the MCP launch post, which cover privacy and third-party data handling only (0 of 15). Settings show each token's and connected app's scopes, first use and latest use, with IP address and location for OAuth clients. No per-call log was found (5 of 15). security.txt valid to 24 June 2027, a disclosure policy and a bug bounty paying $100 to $5,000. The privacy help page says Fastmail has not pursued SOC certification, and no ISO 27001 certificate was found (10 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: Fastmail keeps no API changelog, so recency rests on what we could date. The MCP server launched on 22 April 2026, 169 days before this check, and the JMAP section of the developer page carries no date. The MCP help article was last edited on 18 September 2026, which we didn't count as an API change (10 of 30). No dated API entries in the last 90 days. The blog had three posts in that window, on the EU data region, a data recovery and a mailbox view (0 of 20). Support answers by email and form around the clock with a stated 24-hour reply target, and fastmail/JMAP-Samples has one open issue and a last commit on 6 April 2026. No developer forum was found (9 of 15 for a closed service). No official SDK. We couldn't reach the MCP registry to look for a Fastmail entry (3 of 15). The samples repository takes dependency updates and has no CI (4 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. Agent ergonomics, 73 out of 100, up to 4.4 more on the total

Why it scored 73: Graded on the JMAP API as RFC 8620 and 8621 specify it. Every `/get` call takes a `properties` list, `Email/get` returns no body text unless `fetchTextBodyValues` is set, `maxBodyValueBytes` caps it, and method calls batch into one request with back-references, so a search and a fetch cost one round trip. The MCP tool count was not read (22 of 25). `/query` takes a structured filter, sort, `position`, `anchor`, `limit` and `collapseThreads`, with `/changes` and `/queryChanges` for sync (20). Errors are typed at three levels, request problem details, method errors such as `invalidArguments` and per-object SetErrors such as `tooManyRecipients` and `forbiddenFrom`. Fastmail adds no error guide of its own (17 of 20). `ifInState` makes a write conditional on the state the client last saw, and creation IDs tie results to requests, but there is no idempotency key for sending. MCP annotations were not read (9 of 20). Few arguments are required and defaults are stated, but there is no official SDK in any language, and a send takes an Email object, an Identity and an EmailSubmission (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.

## 7. Transparency & trust, 74 out of 100, up to 2.3 more on the total

Made of editorial 62, provenance 86.

Why it scored 74: The editorial half. A closed service with API Terms of Service and an API Developer Policy that state the grant, the limits and the governing law (Victoria, Australia). The protocol is an open IETF standard and the sample code is MIT (15 of 30). The privacy policy, updated 2 July 2026, names retention periods (IP logs up to one year, the address up to twelve months after closure, account data deleted within seven days of closure) and says customer data is not used to train AI. The DPA is public in five sections and three annexes, a transparency report counts law-enforcement requests by year, and the data residency page says backups, logs and debug systems stay in the US for EU accounts. No retention period for backups beyond the seven-day minimum was found (26 of 30). No deprecation policy. The API terms say Fastmail may make backwards-incompatible changes and will try to give notice (4 of 20). DPA Annex 3 names 36 subprocessors with the purpose of each, without a country for each, and the security help page names data centres in Philadelphia, St. Louis and Amsterdam (17 of 20). The provenance half is computed from the provenance block.

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)
- Privacy policy: read, states 7 of the 8 things a reader expects (9.3 of 10)
- Changelog: 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 JMAP session object, with Fastmail's request limits and capability list. It needs a signed-in account and we made no authenticated call
- unchecked: the MCP server's tool list, tool count and annotations, which sit behind OAuth or a token
- unchecked: whether Fastmail's MCP server is in the official MCP registry. registry.modelcontextprotocol.io timed out on 8 October 2026
- unchecked: whether API tokens can be created during the free trial. The help centre says only that Basic plans can't create them
- The developer page says OAuth clients are registered by hand through the partnerships team, while the authorisation server metadata lists a `registration_endpoint` at `/oauth/register`. We did not call it
- The developer page lists `urn:ietf:params:jmap:*` scopes, while the authorisation server metadata lists `urn:ietf:params:oauth:scope:mail`, contacts and calendars and the MCP scope. Which set a new client gets is not established
- How the customer terms' ban on programmatically generated email to outside addresses applies to an agent sending on its owner's behalf through the send scope Fastmail itself sells
- Whether the three partial outages touched JMAP. The status page has no JMAP or API component
- No API changelog was found. The API terms refer to changelogs without a link
- The 10 points for the incident record read three partial outages of one to eight hours, each for a subset of users, as one major outage. A stricter reading gives 0
- The 10 points for recency count the MCP launch of 22 April 2026. Counting the 18 September 2026 edit of the MCP help article would give 30
- www.fastmail.com/robots.txt carries the content-signal preamble as comments with no `Content-Signal` directive and no user-agent rules on 8 October 2026, so nothing was left unread on that ground
- In April 2026 a bug in account moves wiped folders in 8 accounts. Fastmail says all mail was restored from backups and disclosed it on 19 August 2026. We took no deduction, since it is not a security incident and was fixed and documented
- No performance, sync delay or send-limit figure was measured

## Weaknesses

- The customer terms forbid programmatically generated email to addresses outside the account, and say the service is not for machine-to-machine workflows
- No OpenAPI file, llms.txt, official SDK or API changelog. The developer page points to the RFCs and four sample scripts
- No request rate limit, Retry-After guidance or SLA was found, and the API terms allow backwards-incompatible changes with notice only promised as an attempt
- Three partial outages between 25 August and 7 October 2026 blocked access for some users for 8 hours 16 minutes, 1 hour 8 minutes and 5 hours 23 minutes
- API tokens are not available on Basic plans, and the developer page says OAuth clients are registered by hand through the partnerships team
- No idempotency key on `EmailSubmission/set`, no confirmation step for delete or send, and no per-call access log
- Fastmail says it has not pursued SOC certification, and no ISO 27001 certificate 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.

- Fetch `https://api.fastmail.com/jmap/session` with `Authorization: Bearer <token>` first. It returns the API URL, account IDs and the request limits to stay under
- Ask the owner for a read-only token unless the task writes. Sending needs both the Email and Email submission scopes
- Request only the `properties` you need. Body text is not returned unless `fetchTextBodyValues` is set, and `maxBodyValueBytes` caps it
- Sync with `Email/changes` from a stored state. On `cannotCalculateChanges`, fetch again from scratch
- Do not send generated mail to outside recipients without the owner's review. The customer terms forbid it, and trial accounts stop at 120 messages a day

## 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.