{
  "fixes": {
    "slug": "photon",
    "name": "Photon",
    "listing": "https://www.anchorterminal.com/tools/photon",
    "markdown": "# Fix list: Photon\n\nFrom Anchor Terminal's listing at https://www.anchorterminal.com/tools/photon, the October 2026 research run, assessed 16 October 2026. Grade D, 53.5 out of 100.\n\nThis 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.\n\nFor a coding agent working on Photon: 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.\n\n## 1. Security \u0026 auth, 29 out of 100, up to 12.4 more on the total\n\nWhy it scored 29: Project ID and secret over HTTP Basic for the Spectrum API. The secret grants full management access to the project, doesn't expire on its own and is rotated with `photon projects regenerate-secret`. The SDK mints per-line iMessage tokens it renews at 80 per cent of their lifetime, and the Dashboard API takes OAuth 2.1 with PKCE and consented scopes. No secret in a URL is documented (17 of 30). No read-only project credential. The CLI asks before removing a user (3 of 20). Inbound messages come from anyone who texts a line, and no prompt-injection guidance was found in the Stable docs (0 of 15). OpenTelemetry tracing is opt-in and per-line analytics are in the dashboard. API logs and traces arrive with Photon 2.0, not Stable (4 of 15). No security.txt on photon.codes. The API clients' repository has a SECURITY.md with private vulnerability reporting, and the privacy policy mentions regular penetration testing. No bug bounty or certification was found (5 of 20). 29.\n\nThe checklist (https://www.anchorterminal.com/benchmark/#checklist-security):\n\n- 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.\n- 0 to 20, read-only or least-privilege modes, and confirmation or approval for destructive actions.\n- 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.\n- 0 to 15, audit logs or per-call visibility for the operator.\n- 0 to 20, a security programme. security.txt or a disclosure policy, a bug bounty, SOC 2 or ISO 27001, advisories handled in public.\n\nModels 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.\n\n## 2. Payments \u0026 pricing, 20 out of 100, up to 10 more on the total\n\nWhy it scored 20: No x402, MPP or L402 (0). Plan prices are public. Free at $0 for up to 10 users, Pro at $25 a month for up to 100, Business at $250 per dedicated line a month and Enterprise on request, with no per-message fee for iMessage (10 of 20). A $0 Free plan exists, but whether it needs a card wasn't established, and the terms limit use 'on a free or trial basis' to internal evaluation (10 of 20). A person signs up in the dashboard, and the CLI's device login needs approval in a browser (0). 20.\n\nThe checklist (https://www.anchorterminal.com/benchmark/#checklist-payments):\n\nThe published rubric, also on the [x402 page](https://www.anchorterminal.com/x402/).\n\n- 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.\n- 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.\n- 20, a free tier or trial that doesn't need a card.\n- 20, autonomous onboarding, meaning an agent can get access without a person signing up in a browser (keyless use, x402, a programmatic key API).\n\nPayment 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.\n\nOpen-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.\n\n## 3. Reliability, 68 out of 100, up to 6.4 more on the total\n\nWhy it scored 68: Read with the hosted lines for Spectrum Cloud, the managed lines the SDK connects to. status.photon.codes runs on incident.io with components for the Spectrum API, webhook delivery, billing, each iMessage plan and the Mac nodes (20). Since 18 July 2026 it records one full outage, iMessage on every plan for 2 hours 7 minutes on 2 September after a power failure at Photon's data centre. It also records an 8-minute partial Spectrum API outage on 7 August (out of memory), a 9-minute error spike on 22 August, about an hour of failed RCS sends on 9 October, and over 200 automated 'macOS clusters were encountering degraded performance' entries, one lasting about two hours on 16 September (10 of 30). Rate limits with numbers, 5 requests a second per project on the Spectrum API, 5,000 messages per server a day and 50 new conversations per line a day (15). The API docs say 5xx is safe to retry with backoff, and the recovery guide gives stable `clientGuid` values for deduplicating retried sends. No `Retry-After` or 429 guidance is documented for the Spectrum API (10 of 15). No SLA document. The pricing page names a contractual SLA only for Enterprise, and the terms disclaim delivery and uninterrupted service (3 of 10). The Stable product is generally available (10). 68.\n\nThe checklist (https://www.anchorterminal.com/benchmark/#checklist-reliability):\n\nHosted APIs, MCP servers, models and platforms.\n\n- 20, a public status page with component history (Statuspage, Instatus, BetterStack or the vendor's own).\n- 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.\n- 15, rate limits documented with numbers.\n- 15, documented 429 or overload handling (Retry-After, backoff guidance), and idempotency keys or safe-retry guidance where writes are involved.\n- 10, an SLA published for any paid tier.\n- 10, the surface agents use is generally available, not beta or preview.\n\nLocal packages, SDKs, frameworks and stdio MCP servers.\n\n- 20, installs from an official package with supported runtimes stated.\n- 25, a public CI and test suite, passing on the default branch.\n- 0 to 25, open crash or regression issues relative to activity (25 for few and handled, 0 for many, old and unanswered).\n- 15, semver discipline and breaking changes called out in a changelog.\n- 15, version 1.0 or later, or declared stable.\n\nProtocols are read from their reference implementations, the public facilitators or servers, spec stability and test vectors.\n\n## 4. Agent ergonomics, 63 out of 100, up to 6 more on the total\n\nWhy it scored 63: Read for the TypeScript SDK an agent server runs. One async message stream for every provider, one send call per space and environment fallbacks keep the code small. Management responses can't be sized beyond list limits (17 of 25). The user and template lists take `limit` with `offset` or cursors and filters, while the other lists take none (12 of 20). API errors return a status code with `{ succeed, message }`, `SpectrumCloudError` carries `status` and `code`, and unsupported operations throw `UnsupportedError`. No error-code catalogue for Stable was found (13 of 20). The recovery guide has callers set stable `clientGuid` values that the transport deduplicates, the advanced iMessage kit takes a `clientMessageId`, and webhooks are at-least-once with dedupe guidance. Spectrum API writes take no idempotency key (13 of 20). `imessage.config()` with no arguments discovers lines and renews tokens, but Stable has one official SDK language, TypeScript (8 of 15). 63.\n\nThe checklist (https://www.anchorterminal.com/benchmark/#checklist-ergonomics):\n\n- 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).\n- 20, pagination, filtering and output-size controls.\n- 20, actionable, documented error responses, codes and messages an agent can recover from.\n- 20, idempotency or safe retries, and for MCP the `readOnlyHint` and `destructiveHint` annotations.\n- 15, sensible defaults, few required parameters, and official SDKs in at least two languages.\n\nModels 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.\n\n## 5. Schema \u0026 documentation, 75 out of 100, up to 4.1 more on the total\n\nWhy it scored 75: Read for the Stable surface. Public OpenAPI 3.1 documents for the Spectrum API (54 operations) and the Dashboard API, and TypeScript declarations for the SDK. Sending runs over gRPC inside the SDK with no public contract, and the Spectrum OpenAPI declares no security scheme (18 of 25). llms.txt, llms-full.txt, a Markdown twin of every page and a documentation MCP server (10). The docs say when to use the API (management) and when the SDK (runtime messaging), and the deliverability guide lists what to do and avoid (16 of 20). Config is validated with zod, spaces and services carry enums, and the OpenAPI request bodies are typed (12 of 15). TypeScript examples on every page, signature verifiers in four languages, a response code table for the API and documented SDK error classes (13 of 15). spectrum-ts follows semver and reached 12.11.0 after 79 npm versions in six months, but no Stable changelog was readable. The weekly updates stop at 20 June 2026, and GitHub release pages were refused by our session's proxy (6 of 15). 75.\n\nThe checklist (https://www.anchorterminal.com/benchmark/#checklist-schema):\n\nAPIs and MCP servers.\n\n- 25, a machine-readable contract (a public OpenAPI file or similar; for MCP, typed JSON Schema inputs on every tool).\n- 10, llms.txt or Markdown docs served for agents.\n- 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.\n- 0 to 15, typed inputs with enums, constraints and required fields, and no free-form JSON blobs.\n- 0 to 15, examples and documented error responses.\n- 15, versioning and a public changelog.\n\nModels 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.\n\n## 6. Transparency \u0026 trust, 58 out of 100, up to 3.7 more on the total\n\nMade of editorial 38, provenance 78.\n\nWhy it scored 58: spectrum-ts, its providers, imessage-kit and the CLI are MIT, while Spectrum Cloud and the managed lines are closed. The undated terms still carry template placeholders, such as '[California]' for governing law and a bracketed acceptable use policy that links to a deliverability guide (18 of 30). The privacy policy (25 September 2026) covers account data and deletion within 30 days of closing an account, but says nothing about message content or end users' data. The terms let Photon use de-identified Usage Data to train its models. No DPA or retention period for messages was found (8 of 30). No deprecation policy. Migration notes for removed SDK options carry no dates (4 of 20). SDK telemetry is opt-in and documented with its OTLP endpoint. Hosting is described as Photon's own data centre alongside cloud providers, with no sub-processor list or regions (8 of 20). 38.\n\nThe checklist (https://www.anchorterminal.com/benchmark/#checklist-transparency):\n\n- 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.\n- 0 to 30, data handling and retention statements that agree with each other (privacy policy, DPA, retention periods, subprocessors).\n- 0 to 20, a deprecation policy or notices with dates.\n- 0 to 20, telemetry disclosed with an opt-out (local software), or subprocessors and data locations disclosed (hosted).\n\nThe other half of Transparency and trust is the provenance score, computed from checked facts (below). The category score is the mean of the two.\n\nProvenance checks not met in full (half of this category, computed from checked facts):\n\n- Domain age: photon.codes, registered 2025-10-10 (1 year) (3 of 15)\n- security.txt: not found (0 of 10)\n\n## 7. Maintenance \u0026 community, 78 out of 100, up to 1.9 more on the total\n\nWhy it scored 78: spectrum-ts 12.11.0 on 15 October 2026 (30). 14 tagged releases between 20 July and 15 October 2026 (20). The SDK is open source, but GitHub issue pages were refused by our session's proxy. Support is by email and Discord, and Photon says its team works around the clock. The weekly updates stopped after 20 June 2026 (8 of 25). The official TypeScript SDK and CLI are current, and the Beta clients cover four languages. No messaging MCP server, so no registry entry (12 of 15). CI runs lint, type checks, an artefact check and the test suite on Node and Bun, with pinned actions and Renovate. Whether it passes on main wasn't checked (8 of 10). 78.\n\nThe checklist (https://www.anchorterminal.com/benchmark/#checklist-maintenance):\n\n- 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.\n- 20, at least three releases or dated changelog entries in the last 90 days.\n- 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.\n- 15, presence in the official MCP registry under a verified namespace (MCP servers), or current official SDKs (APIs and models).\n- 10, package health, current dependencies and CI.\n\nModels are read for deprecation notice periods and model churn rather than release counts.\n\n## Deductions\n\nEach comes off the total. A fixed and documented problem counts for less at the next check.\n\n- 2026-10-16. The iMessage platform page states '99.95% Uptime' and infrastructure 'built for speed, stability, and 99.999% uptime' (https://photon.codes/platform/imessage). Photon's own status page shows 99.89 per cent for each iMessage component from July to October 2026, including a 2-hour 7-minute full outage on 2 September 2026 (https://status.photon.codes, https://status.photon.codes/incidents/ry86mwnd). A misleading claim, -2.\n\n## What we couldn't check\n\nWhat 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.\n\n- unchecked: GitHub stars, open issues and release pages for photon-hq repositories, because github.com web pages and the GitHub API were refused by our session's proxy. Tags, CI and source were read from shallow clones\n- unchecked: whether the Free plan, Stable or Beta, needs a payment method at sign-up. The Beta billing page says a payment method is needed before paid usage\n- unchecked: the terms of the $100 credit. The announcement says '$100 in free credit to start building', and photon.codes/beta and the docs give no amount, expiry or eligibility\n- unchecked: whether CI passes on spectrum-ts main\n- WebFetch was refused for https://photon.codes/docs/llms.txt (permission request not answered), so every Photon page in this dossier was read with curl, after robots.txt\n- The terms' Trials and Betas section covers access 'on a free or trial basis'. Whether Photon applies its 30-day internal-evaluation limit to the Free plan isn't stated\n- How Photon sends iMessage beyond the announcement's macOS clusters and Apple Account registration isn't documented, and neither the docs nor the terms state Apple's position on it\n- The Beta get-started page says the API clients' messaging methods aren't released yet, while the clients' changelog added them in 0.4.2 on 15 October 2026\n- The deduction of 2 for the uptime claims is a judgement call for the desk. The page says infrastructure is 'built for speed, stability, and 99.999% uptime', which can be read as a design target, but its '99.95% Uptime' figure is above the 99.89 per cent its status page shows\n- The status page doesn't state the time zone of its timestamps. Durations were taken from its displayed start and resolve times\n- photon.codes shows '4.6k' beside its GitHub link. We couldn't read the GitHub count it refers to\n\n## Weaknesses\n\n- The status page logged a 2-hour full iMessage outage on 2 September 2026, from a power failure at Photon's data centre\n- The project secret grants full management access, never expires on its own and has no read-only variant\n- The terms limit free, trial and beta use to internal evaluation, 30 days by default, with liability capped at US$50\n- No SLA document. Only the Enterprise plan mentions a contractual SLA, and the terms disclaim delivery\n- No prompt-injection guidance for inbound messages, and no DPA, sub-processor list or message retention period was found\n\n## What costs an agent a turn today\n\nThe notes we give agents before they call it. Each one is a workaround an agent shouldn't need.\n\n- Use the Stable docs and `spectrum-ts` unless the project ID starts with `pho_prj_`. Stable and Beta packages, hosts and credentials don't mix\n- Design flows so the person texts first. Photon ties line flags to cold outreach, bursts and more than two or three follow-ups\n- Stay under 5,000 messages per server a day and 50 new conversations per line a day. Further sends are rejected until the window resets\n- Give each outbound message a stable `clientGuid` so a retried job doesn't send it twice\n- Dedupe webhooks on `X-Spectrum-Webhook-Id` plus the event's own id, answer within 30 seconds, and reconcile, because events are dropped after six attempts\n- Keep the project secret out of the agent's context. It grants full management access and doesn't expire\n\n## What the review panel asked for\n\n- Programmatic project creation\n- State Free plan card needs\n- Publish an error-code catalogue\n- Declare the security scheme in OpenAPI\n- Document Retry-After on 429\n- Publish an SLA for all paid plans\n- read-only project credential\n- prompt-injection guidance for inbound\n\n## When it's done\n\nSend 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.\n",
    "grade": "D",
    "score": 53.5,
    "assessed": "2026-10-16",
    "run": "October 2026 research run",
    "categories": [
      {
        "key": "security",
        "name": "Security \u0026 auth",
        "score": 29,
        "maxGain": 12.4,
        "reason": "Project ID and secret over HTTP Basic for the Spectrum API. The secret grants full management access to the project, doesn't expire on its own and is rotated with `photon projects regenerate-secret`. The SDK mints per-line iMessage tokens it renews at 80 per cent of their lifetime, and the Dashboard API takes OAuth 2.1 with PKCE and consented scopes. No secret in a URL is documented (17 of 30). No read-only project credential. The CLI asks before removing a user (3 of 20). Inbound messages come from anyone who texts a line, and no prompt-injection guidance was found in the Stable docs (0 of 15). OpenTelemetry tracing is opt-in and per-line analytics are in the dashboard. API logs and traces arrive with Photon 2.0, not Stable (4 of 15). No security.txt on photon.codes. The API clients' repository has a SECURITY.md with private vulnerability reporting, and the privacy policy mentions regular penetration testing. No bug bounty or certification was found (5 of 20). 29.",
        "checklist": [
          "- 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.\n- 0 to 20, read-only or least-privilege modes, and confirmation or approval for destructive actions.\n- 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.\n- 0 to 15, audit logs or per-call visibility for the operator.\n- 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."
        ],
        "checklistUrl": "https://www.anchorterminal.com/benchmark/#checklist-security"
      },
      {
        "key": "payments",
        "name": "Payments \u0026 pricing",
        "score": 20,
        "maxGain": 10,
        "reason": "No x402, MPP or L402 (0). Plan prices are public. Free at $0 for up to 10 users, Pro at $25 a month for up to 100, Business at $250 per dedicated line a month and Enterprise on request, with no per-message fee for iMessage (10 of 20). A $0 Free plan exists, but whether it needs a card wasn't established, and the terms limit use 'on a free or trial basis' to internal evaluation (10 of 20). A person signs up in the dashboard, and the CLI's device login needs approval in a browser (0). 20.",
        "checklist": [
          "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.\n- 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.\n- 20, a free tier or trial that doesn't need a card.\n- 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."
        ],
        "checklistUrl": "https://www.anchorterminal.com/benchmark/#checklist-payments"
      },
      {
        "key": "reliability",
        "name": "Reliability",
        "score": 68,
        "maxGain": 6.4,
        "reason": "Read with the hosted lines for Spectrum Cloud, the managed lines the SDK connects to. status.photon.codes runs on incident.io with components for the Spectrum API, webhook delivery, billing, each iMessage plan and the Mac nodes (20). Since 18 July 2026 it records one full outage, iMessage on every plan for 2 hours 7 minutes on 2 September after a power failure at Photon's data centre. It also records an 8-minute partial Spectrum API outage on 7 August (out of memory), a 9-minute error spike on 22 August, about an hour of failed RCS sends on 9 October, and over 200 automated 'macOS clusters were encountering degraded performance' entries, one lasting about two hours on 16 September (10 of 30). Rate limits with numbers, 5 requests a second per project on the Spectrum API, 5,000 messages per server a day and 50 new conversations per line a day (15). The API docs say 5xx is safe to retry with backoff, and the recovery guide gives stable `clientGuid` values for deduplicating retried sends. No `Retry-After` or 429 guidance is documented for the Spectrum API (10 of 15). No SLA document. The pricing page names a contractual SLA only for Enterprise, and the terms disclaim delivery and uninterrupted service (3 of 10). The Stable product is generally available (10). 68.",
        "checklist": [
          "Hosted APIs, MCP servers, models and platforms.",
          "- 20, a public status page with component history (Statuspage, Instatus, BetterStack or the vendor's own).\n- 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.\n- 15, rate limits documented with numbers.\n- 15, documented 429 or overload handling (Retry-After, backoff guidance), and idempotency keys or safe-retry guidance where writes are involved.\n- 10, an SLA published for any paid tier.\n- 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.\n- 25, a public CI and test suite, passing on the default branch.\n- 0 to 25, open crash or regression issues relative to activity (25 for few and handled, 0 for many, old and unanswered).\n- 15, semver discipline and breaking changes called out in a changelog.\n- 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."
        ],
        "checklistUrl": "https://www.anchorterminal.com/benchmark/#checklist-reliability"
      },
      {
        "key": "ergonomics",
        "name": "Agent ergonomics",
        "score": 63,
        "maxGain": 6,
        "reason": "Read for the TypeScript SDK an agent server runs. One async message stream for every provider, one send call per space and environment fallbacks keep the code small. Management responses can't be sized beyond list limits (17 of 25). The user and template lists take `limit` with `offset` or cursors and filters, while the other lists take none (12 of 20). API errors return a status code with `{ succeed, message }`, `SpectrumCloudError` carries `status` and `code`, and unsupported operations throw `UnsupportedError`. No error-code catalogue for Stable was found (13 of 20). The recovery guide has callers set stable `clientGuid` values that the transport deduplicates, the advanced iMessage kit takes a `clientMessageId`, and webhooks are at-least-once with dedupe guidance. Spectrum API writes take no idempotency key (13 of 20). `imessage.config()` with no arguments discovers lines and renews tokens, but Stable has one official SDK language, TypeScript (8 of 15). 63.",
        "checklist": [
          "- 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).\n- 20, pagination, filtering and output-size controls.\n- 20, actionable, documented error responses, codes and messages an agent can recover from.\n- 20, idempotency or safe retries, and for MCP the `readOnlyHint` and `destructiveHint` annotations.\n- 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."
        ],
        "checklistUrl": "https://www.anchorterminal.com/benchmark/#checklist-ergonomics"
      },
      {
        "key": "schema",
        "name": "Schema \u0026 documentation",
        "score": 75,
        "maxGain": 4.1,
        "reason": "Read for the Stable surface. Public OpenAPI 3.1 documents for the Spectrum API (54 operations) and the Dashboard API, and TypeScript declarations for the SDK. Sending runs over gRPC inside the SDK with no public contract, and the Spectrum OpenAPI declares no security scheme (18 of 25). llms.txt, llms-full.txt, a Markdown twin of every page and a documentation MCP server (10). The docs say when to use the API (management) and when the SDK (runtime messaging), and the deliverability guide lists what to do and avoid (16 of 20). Config is validated with zod, spaces and services carry enums, and the OpenAPI request bodies are typed (12 of 15). TypeScript examples on every page, signature verifiers in four languages, a response code table for the API and documented SDK error classes (13 of 15). spectrum-ts follows semver and reached 12.11.0 after 79 npm versions in six months, but no Stable changelog was readable. The weekly updates stop at 20 June 2026, and GitHub release pages were refused by our session's proxy (6 of 15). 75.",
        "checklist": [
          "APIs and MCP servers.",
          "- 25, a machine-readable contract (a public OpenAPI file or similar; for MCP, typed JSON Schema inputs on every tool).\n- 10, llms.txt or Markdown docs served for agents.\n- 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.\n- 0 to 15, typed inputs with enums, constraints and required fields, and no free-form JSON blobs.\n- 0 to 15, examples and documented error responses.\n- 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."
        ],
        "checklistUrl": "https://www.anchorterminal.com/benchmark/#checklist-schema"
      },
      {
        "key": "transparency",
        "name": "Transparency \u0026 trust",
        "score": 58,
        "maxGain": 3.7,
        "reason": "spectrum-ts, its providers, imessage-kit and the CLI are MIT, while Spectrum Cloud and the managed lines are closed. The undated terms still carry template placeholders, such as '[California]' for governing law and a bracketed acceptable use policy that links to a deliverability guide (18 of 30). The privacy policy (25 September 2026) covers account data and deletion within 30 days of closing an account, but says nothing about message content or end users' data. The terms let Photon use de-identified Usage Data to train its models. No DPA or retention period for messages was found (8 of 30). No deprecation policy. Migration notes for removed SDK options carry no dates (4 of 20). SDK telemetry is opt-in and documented with its OTLP endpoint. Hosting is described as Photon's own data centre alongside cloud providers, with no sub-processor list or regions (8 of 20). 38.",
        "blend": "editorial 38, provenance 78",
        "checklist": [
          "- 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.\n- 0 to 30, data handling and retention statements that agree with each other (privacy policy, DPA, retention periods, subprocessors).\n- 0 to 20, a deprecation policy or notices with dates.\n- 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."
        ],
        "checklistUrl": "https://www.anchorterminal.com/benchmark/#checklist-transparency"
      },
      {
        "key": "maintenance",
        "name": "Maintenance \u0026 community",
        "score": 78,
        "maxGain": 1.9,
        "reason": "spectrum-ts 12.11.0 on 15 October 2026 (30). 14 tagged releases between 20 July and 15 October 2026 (20). The SDK is open source, but GitHub issue pages were refused by our session's proxy. Support is by email and Discord, and Photon says its team works around the clock. The weekly updates stopped after 20 June 2026 (8 of 25). The official TypeScript SDK and CLI are current, and the Beta clients cover four languages. No messaging MCP server, so no registry entry (12 of 15). CI runs lint, type checks, an artefact check and the test suite on Node and Bun, with pinned actions and Renovate. Whether it passes on main wasn't checked (8 of 10). 78.",
        "checklist": [
          "- 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.\n- 20, at least three releases or dated changelog entries in the last 90 days.\n- 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.\n- 15, presence in the official MCP registry under a verified namespace (MCP servers), or current official SDKs (APIs and models).\n- 10, package health, current dependencies and CI.",
          "Models are read for deprecation notice periods and model churn rather than release counts."
        ],
        "checklistUrl": "https://www.anchorterminal.com/benchmark/#checklist-maintenance"
      }
    ],
    "provenance": [
      {
        "label": "Domain age",
        "value": "photon.codes, registered 2025-10-10 (1 year)",
        "points": 3,
        "max": 15
      },
      {
        "label": "security.txt",
        "value": "not found",
        "points": 0,
        "max": 10
      }
    ],
    "deductions": [
      "2026-10-16. The iMessage platform page states '99.95% Uptime' and infrastructure 'built for speed, stability, and 99.999% uptime' (https://photon.codes/platform/imessage). Photon's own status page shows 99.89 per cent for each iMessage component from July to October 2026, including a 2-hour 7-minute full outage on 2 September 2026 (https://status.photon.codes, https://status.photon.codes/incidents/ry86mwnd). A misleading claim, -2."
    ],
    "unchecked": [
      "unchecked: GitHub stars, open issues and release pages for photon-hq repositories, because github.com web pages and the GitHub API were refused by our session's proxy. Tags, CI and source were read from shallow clones",
      "unchecked: whether the Free plan, Stable or Beta, needs a payment method at sign-up. The Beta billing page says a payment method is needed before paid usage",
      "unchecked: the terms of the $100 credit. The announcement says '$100 in free credit to start building', and photon.codes/beta and the docs give no amount, expiry or eligibility",
      "unchecked: whether CI passes on spectrum-ts main",
      "WebFetch was refused for https://photon.codes/docs/llms.txt (permission request not answered), so every Photon page in this dossier was read with curl, after robots.txt",
      "The terms' Trials and Betas section covers access 'on a free or trial basis'. Whether Photon applies its 30-day internal-evaluation limit to the Free plan isn't stated",
      "How Photon sends iMessage beyond the announcement's macOS clusters and Apple Account registration isn't documented, and neither the docs nor the terms state Apple's position on it",
      "The Beta get-started page says the API clients' messaging methods aren't released yet, while the clients' changelog added them in 0.4.2 on 15 October 2026",
      "The deduction of 2 for the uptime claims is a judgement call for the desk. The page says infrastructure is 'built for speed, stability, and 99.999% uptime', which can be read as a design target, but its '99.95% Uptime' figure is above the 99.89 per cent its status page shows",
      "The status page doesn't state the time zone of its timestamps. Durations were taken from its displayed start and resolve times",
      "photon.codes shows '4.6k' beside its GitHub link. We couldn't read the GitHub count it refers to"
    ],
    "weaknesses": [
      "The status page logged a 2-hour full iMessage outage on 2 September 2026, from a power failure at Photon's data centre",
      "The project secret grants full management access, never expires on its own and has no read-only variant",
      "The terms limit free, trial and beta use to internal evaluation, 30 days by default, with liability capped at US$50",
      "No SLA document. Only the Enterprise plan mentions a contractual SLA, and the terms disclaim delivery",
      "No prompt-injection guidance for inbound messages, and no DPA, sub-processor list or message retention period was found"
    ],
    "agentNotes": [
      "Use the Stable docs and `spectrum-ts` unless the project ID starts with `pho_prj_`. Stable and Beta packages, hosts and credentials don't mix",
      "Design flows so the person texts first. Photon ties line flags to cold outreach, bursts and more than two or three follow-ups",
      "Stay under 5,000 messages per server a day and 50 new conversations per line a day. Further sends are rejected until the window resets",
      "Give each outbound message a stable `clientGuid` so a retried job doesn't send it twice",
      "Dedupe webhooks on `X-Spectrum-Webhook-Id` plus the event's own id, answer within 30 seconds, and reconcile, because events are dropped after six attempts",
      "Keep the project secret out of the agent's context. It grants full management access and doesn't expire"
    ],
    "requests": [
      {
        "text": "Programmatic project creation",
        "reviews": 1
      },
      {
        "text": "State Free plan card needs",
        "reviews": 1
      },
      {
        "text": "Publish an error-code catalogue",
        "reviews": 1
      },
      {
        "text": "Declare the security scheme in OpenAPI",
        "reviews": 1
      },
      {
        "text": "Document Retry-After on 429",
        "reviews": 1
      },
      {
        "text": "Publish an SLA for all paid plans",
        "reviews": 1
      },
      {
        "text": "read-only project credential",
        "reviews": 1
      },
      {
        "text": "prompt-injection guidance for inbound",
        "reviews": 1
      }
    ],
    "recheck": "https://www.anchorterminal.com/builders/#disputes"
  },
  "meta": {
    "attribution": "Anchor Terminal (https://www.anchorterminal.com)",
    "attributionRequired": true,
    "credit": "CC BY 4.0, credit required: when you show, quote or pass on anything from here (to a user in an answer, in a product, a report or a dataset), say it's from Anchor Terminal (anchorterminal.com) and link the page or API address you took it from. Reviews also credit their author.",
    "docs": "https://www.anchorterminal.com/docs/",
    "generatedAt": "2026-10-16",
    "license": "CC-BY-4.0",
    "method": "https://www.anchorterminal.com/benchmark/",
    "methodology": "0.4",
    "openapi": "https://www.anchorterminal.com/openapi.json",
    "preview": false,
    "run": "2026-10-01",
    "runLabel": "October 2026 research run",
    "terms": "https://www.anchorterminal.com/terms/#data"
  }
}