{
  "fixes": {
    "slug": "persona",
    "name": "Persona",
    "listing": "https://www.anchorterminal.com/tools/persona",
    "markdown": "# Fix list: Persona\n\nFrom Anchor Terminal's listing at https://www.anchorterminal.com/tools/persona, the October 2026 research run, assessed 8 October 2026. Grade B, 69.5 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 Persona: 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. Payments \u0026 pricing, 35 out of 100, up to 8.1 more on the total\n\nWhy it scored 35: Read with the hosted rubric. No x402, MPP or L402 (0). The help centre publishes the Essential plan at $250 a month on an annual contract with 500 services included and $1.50 for each further verification or report. Growth and Enterprise prices come from sales, and the pricing page refused our reader (15 of 20). A 60-day trial with no card gives a sandbox and up to 50 services. The sandbox performs no real verifications (20). A person signs up in a browser with a business email, and production needs a plan and Persona's review of the business (0).\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## 2. Reliability, 70 out of 100, up to 6 more on the total\n\nWhy it scored 70: Read with the hosted lines and scored on the REST API, with the MCP server noted. status.withpersona.com is Persona's own page with ten components and incident history back to October 2024 (20). It lists 11 incidents between 10 July and 8 October 2026, six marked degraded and five partial outage. One is major by our reading, latency and timeouts across all products and the dashboard from 15:30 to 16:40 UTC on 1 September, which Persona traced to Google Cloud us-central1. The others are narrower, among them government ID failures on 30 September (21:58 to 23:12 UTC), document verification failures on 6 October (19:00 to 20:47 UTC) and intermittent OCR errors on 19 and 20 August (10). Limits have numbers, 300 requests a minute per production environment by default, plus creation quotas (15). 429 responses come with `RateLimit-Reset` and `Quota-Reset` headers and backoff guidance starting at 5 seconds, and every POST takes an `Idempotency-Key` stored for at least 24 hours (15). No SLA was found in the docs or help centre, and the terms and pricing page refused our reader (0). The API is generally available under dated versions. The MCP server isn't labelled beta, though OAuth for it is still to come (10).\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## 3. Security \u0026 auth, 73 out of 100, up to 4.7 more on the total\n\nWhy it scored 73: Bearer API keys per environment, each limited to about 50 read and write permissions per resource, with an IP allowlist, a per-key rate limit, an expiry endpoint and written rotation steps. A new key has every standard permission until someone limits it, and key management permissions are off by default (27 of 30). A read-only key is possible and Persona recommends one dedicated key per agent. Redaction is irreversible with no confirmation step, the MCP server exposes all 190 tools whatever the key allows, and payload filters that restrict visible records are Enterprise only. An investigation lock can block redaction of a verification (12 of 20). Responses carry end-user input, document extractions, adverse media and website content. The MCP article asks customers to check their own security requirements and to use a read-only key, with no prompt-injection guidance found (4 of 15). API logs for two weeks and user audit logs for six months, both through the API, and a `Request-Id` on every response (15). security.txt valid until 17 February 2027, a disclosure form run on Bugcrowd, and SOC 2 Type II and ISO 27001 per the help centre. No public advisories or bounty terms were found, and the trust centre needs JavaScript we couldn't read (15 of 20).\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## 4. Agent ergonomics, 74 out of 100, up to 4.2 more on the total\n\nWhy it scored 74: Scored on the REST API. `fields` is accepted on 213 operations and `include` on 181, so a caller can drop related objects and attributes. Default inquiry responses in JSON API form are large, and the MCP server lists 190 tools in about 117 KB with no toolsets and without `fields` or `include` (20 of 25). Cursor pagination with `page[size]` from 1 to 100, filters on 17 list operations, and search endpoints for inquiries, accounts, cases and webhook events (18 of 20). Errors are an HTTP status with `title` and `details` such as \"data.attributes.name is missing\", with advice per status code. There's no list of machine-readable error codes (14 of 20). `Idempotency-Key` on every POST with parameter checking. The MCP tools carry no readOnlyHint or destructiveHint and don't expose the idempotency header (16 of 20). No official server-side SDK was found. The npm, iOS, Android and React Native packages are clients for the end-user flow, and API examples are raw HTTP in curl, Ruby, Python and JavaScript (6 of 15).\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. Transparency \u0026 trust, 68 out of 100, up to 2.8 more on the total\n\nMade of editorial 46, provenance 90.\n\nWhy it scored 68: The service is closed. The terms of service refused our reader, so we scored licence clarity on what we could read, which is the MIT licence on the npm client and the plan and contract terms in the help centre (10 of 30). The help centre says Persona keeps data until the customer redacts it, that redaction is permanent, that organisation-wide policies run from 2 days to 2 years, and that subject requests are handled in 72 hours. API logs are kept two weeks and audit logs six months. The privacy policy and DPA weren't read (18 of 30). The versioning page defines breaking and non-breaking changes and keeps dated versions back to 2020-05-18, with no stated end date for old versions. On 13 April 2026 the changelog recorded, as an improvement, the removal of a backfilled attribute for versions before 2021-08-18 (12 of 20). No sub-processor list was read. The docs name webhook source addresses in Germany, India and the United States and separate European Union sender domains (6 of 20).\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: withpersona.com, registered 2018-09-24 (8 years) (11 of 15)\n- Terms of service: published, but our reader couldn't read it (7 of 10)\n- Privacy policy: published, but our reader couldn't read it (7 of 10)\n\n## 6. Schema \u0026 documentation, 85 out of 100, up to 2.4 more on the total\n\nWhy it scored 85: A public OpenAPI 3.1 file for version 2025-12-08 has 191 paths, 231 operations and 141 webhook definitions, and the MCP server returns typed input schemas for 190 tools (25). llms.txt per docs version and for the help centre, and any page as Markdown by adding `.md` (10). Operation descriptions are short, with a median of 61 characters and 50 of 231 under 40, though guides for each product say when to use inquiries, transactions, verifications and reports (12 of 20). The spec carries 1,070 enums and 2,720 required lists, but the versioning page tells clients not to treat string values as fixed enumerations. In the MCP list 13 tools have an empty input schema, `create-an-api-key` among them, and four contain `$ref` pointers that resolve to nothing in the tool definition (12 of 15). Every operation documents 400, 401, 403 and 429 responses, and guides carry curl requests and sample payloads. The spec file itself has no example values (11 of 15). Dated API versions and a changelog with an RSS feed (15). The two spec links printed at the foot of the versioned llms.txt return 404, and the working file is at /openapi/api-reference.json.\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## 7. Maintenance \u0026 community, 75 out of 100, up to 2.2 more on the total\n\nWhy it scored 75: The latest API changelog entry is 29 September 2026 (30). Four dated entries in the last 90 days, on 15 July, 18 August, 24 August and 29 September (20). A closed service with a changelog and RSS feed, a help centre, a community Slack and a support form in the dashboard. Persona says it can't guarantee support for trial and Startup Programme accounts, and we didn't test a channel (10 of 15). The official packages are clients for the end-user flow. The npm `persona` package reached 5.9.0 on 16 September 2026, and no server-side SDK was found (10 of 15). The npm package is MIT with regular releases. Its source repository and CI aren't linked from the package (5 of 10).\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## 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: withpersona.com marketing pages (home, pricing, security, terms of service, privacy policy, DPA and sub-processor pages) answered our reader with a bot check (HTTP 403), so the terms, the privacy policy, any SLA and any sub-processor list were not read\n- unchecked: the trust centre at trust.withpersona.com, which needs JavaScript, so certification reports and any published sub-processors were not confirmed there\n- unchecked: security incidents, advisories or vendor post-mortems in the last 12 months. Persona's blog and trust pages were unreadable, so the deduction is 0 on what we could read\n- unchecked: whether customers on API versions before 2021-08-18 had notice before the backfilled `extraction-results` attribute was removed on 13 April 2026\n- unchecked: whether the trial sign-up asks for a card (the help centre says it doesn't) and whether a failed verification counts as a billable service\n- unchecked: sandbox rate limits, and whether the MCP server appears in the official MCP registry\n- The versioning page names 2025-12-08 as the latest version, while the `Persona-Version` header on the create-inquiry reference page also lists 2026-09-29\n- The API keys page describes the `mcp.access` permission as access to a Machine Learning Classification Platform, which is unrelated to the Model Context Protocol server\n\n## Weaknesses\n\n- Production access needs Persona's approval and a 12-month plan from $250 a month, and API-only integration is limited to Enterprise plans\n- Eleven incidents on the status page since 10 July 2026, five marked partial outage, most on document and government ID verifications\n- The MCP server lists all 190 tools whatever the key allows, with no annotations, toolsets or read-only subset, and OAuth isn't available yet\n- No official server-side SDK was found. The npm, iOS, Android and React Native packages are clients for the end-user flow\n- The marketing site, terms, privacy policy and pricing page answered our reader with a bot check, and no SLA was found in the docs or help centre\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- Ask for a dedicated key limited to the permissions the task needs. A new key has every standard permission until it's limited\n- Send `Persona-Version` on every call and an `Idempotency-Key` on every POST. A replayed key returns the first result, including a 500\n- Treat redact calls as irreversible. They permanently delete personal data and cascade to downstream objects\n- Stay under 300 requests a minute in production, read `RateLimit-Remaining` and `Quota-Remaining`, and back off on 429 because rejected requests still count\n- Use `fields` and `include` to cut response size. Through MCP these parameters aren't exposed, and only 14 tools take `filter`\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": "B",
    "score": 69.5,
    "assessed": "2026-10-08",
    "run": "October 2026 research run",
    "categories": [
      {
        "key": "payments",
        "name": "Payments \u0026 pricing",
        "score": 35,
        "maxGain": 8.1,
        "reason": "Read with the hosted rubric. No x402, MPP or L402 (0). The help centre publishes the Essential plan at $250 a month on an annual contract with 500 services included and $1.50 for each further verification or report. Growth and Enterprise prices come from sales, and the pricing page refused our reader (15 of 20). A 60-day trial with no card gives a sandbox and up to 50 services. The sandbox performs no real verifications (20). A person signs up in a browser with a business email, and production needs a plan and Persona's review of the business (0).",
        "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": 70,
        "maxGain": 6,
        "reason": "Read with the hosted lines and scored on the REST API, with the MCP server noted. status.withpersona.com is Persona's own page with ten components and incident history back to October 2024 (20). It lists 11 incidents between 10 July and 8 October 2026, six marked degraded and five partial outage. One is major by our reading, latency and timeouts across all products and the dashboard from 15:30 to 16:40 UTC on 1 September, which Persona traced to Google Cloud us-central1. The others are narrower, among them government ID failures on 30 September (21:58 to 23:12 UTC), document verification failures on 6 October (19:00 to 20:47 UTC) and intermittent OCR errors on 19 and 20 August (10). Limits have numbers, 300 requests a minute per production environment by default, plus creation quotas (15). 429 responses come with `RateLimit-Reset` and `Quota-Reset` headers and backoff guidance starting at 5 seconds, and every POST takes an `Idempotency-Key` stored for at least 24 hours (15). No SLA was found in the docs or help centre, and the terms and pricing page refused our reader (0). The API is generally available under dated versions. The MCP server isn't labelled beta, though OAuth for it is still to come (10).",
        "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": "security",
        "name": "Security \u0026 auth",
        "score": 73,
        "maxGain": 4.7,
        "reason": "Bearer API keys per environment, each limited to about 50 read and write permissions per resource, with an IP allowlist, a per-key rate limit, an expiry endpoint and written rotation steps. A new key has every standard permission until someone limits it, and key management permissions are off by default (27 of 30). A read-only key is possible and Persona recommends one dedicated key per agent. Redaction is irreversible with no confirmation step, the MCP server exposes all 190 tools whatever the key allows, and payload filters that restrict visible records are Enterprise only. An investigation lock can block redaction of a verification (12 of 20). Responses carry end-user input, document extractions, adverse media and website content. The MCP article asks customers to check their own security requirements and to use a read-only key, with no prompt-injection guidance found (4 of 15). API logs for two weeks and user audit logs for six months, both through the API, and a `Request-Id` on every response (15). security.txt valid until 17 February 2027, a disclosure form run on Bugcrowd, and SOC 2 Type II and ISO 27001 per the help centre. No public advisories or bounty terms were found, and the trust centre needs JavaScript we couldn't read (15 of 20).",
        "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": "ergonomics",
        "name": "Agent ergonomics",
        "score": 74,
        "maxGain": 4.2,
        "reason": "Scored on the REST API. `fields` is accepted on 213 operations and `include` on 181, so a caller can drop related objects and attributes. Default inquiry responses in JSON API form are large, and the MCP server lists 190 tools in about 117 KB with no toolsets and without `fields` or `include` (20 of 25). Cursor pagination with `page[size]` from 1 to 100, filters on 17 list operations, and search endpoints for inquiries, accounts, cases and webhook events (18 of 20). Errors are an HTTP status with `title` and `details` such as \"data.attributes.name is missing\", with advice per status code. There's no list of machine-readable error codes (14 of 20). `Idempotency-Key` on every POST with parameter checking. The MCP tools carry no readOnlyHint or destructiveHint and don't expose the idempotency header (16 of 20). No official server-side SDK was found. The npm, iOS, Android and React Native packages are clients for the end-user flow, and API examples are raw HTTP in curl, Ruby, Python and JavaScript (6 of 15).",
        "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": "transparency",
        "name": "Transparency \u0026 trust",
        "score": 68,
        "maxGain": 2.8,
        "reason": "The service is closed. The terms of service refused our reader, so we scored licence clarity on what we could read, which is the MIT licence on the npm client and the plan and contract terms in the help centre (10 of 30). The help centre says Persona keeps data until the customer redacts it, that redaction is permanent, that organisation-wide policies run from 2 days to 2 years, and that subject requests are handled in 72 hours. API logs are kept two weeks and audit logs six months. The privacy policy and DPA weren't read (18 of 30). The versioning page defines breaking and non-breaking changes and keeps dated versions back to 2020-05-18, with no stated end date for old versions. On 13 April 2026 the changelog recorded, as an improvement, the removal of a backfilled attribute for versions before 2021-08-18 (12 of 20). No sub-processor list was read. The docs name webhook source addresses in Germany, India and the United States and separate European Union sender domains (6 of 20).",
        "blend": "editorial 46, provenance 90",
        "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": "schema",
        "name": "Schema \u0026 documentation",
        "score": 85,
        "maxGain": 2.4,
        "reason": "A public OpenAPI 3.1 file for version 2025-12-08 has 191 paths, 231 operations and 141 webhook definitions, and the MCP server returns typed input schemas for 190 tools (25). llms.txt per docs version and for the help centre, and any page as Markdown by adding `.md` (10). Operation descriptions are short, with a median of 61 characters and 50 of 231 under 40, though guides for each product say when to use inquiries, transactions, verifications and reports (12 of 20). The spec carries 1,070 enums and 2,720 required lists, but the versioning page tells clients not to treat string values as fixed enumerations. In the MCP list 13 tools have an empty input schema, `create-an-api-key` among them, and four contain `$ref` pointers that resolve to nothing in the tool definition (12 of 15). Every operation documents 400, 401, 403 and 429 responses, and guides carry curl requests and sample payloads. The spec file itself has no example values (11 of 15). Dated API versions and a changelog with an RSS feed (15). The two spec links printed at the foot of the versioned llms.txt return 404, and the working file is at /openapi/api-reference.json.",
        "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": "maintenance",
        "name": "Maintenance \u0026 community",
        "score": 75,
        "maxGain": 2.2,
        "reason": "The latest API changelog entry is 29 September 2026 (30). Four dated entries in the last 90 days, on 15 July, 18 August, 24 August and 29 September (20). A closed service with a changelog and RSS feed, a help centre, a community Slack and a support form in the dashboard. Persona says it can't guarantee support for trial and Startup Programme accounts, and we didn't test a channel (10 of 15). The official packages are clients for the end-user flow. The npm `persona` package reached 5.9.0 on 16 September 2026, and no server-side SDK was found (10 of 15). The npm package is MIT with regular releases. Its source repository and CI aren't linked from the package (5 of 10).",
        "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": "withpersona.com, registered 2018-09-24 (8 years)",
        "points": 11,
        "max": 15
      },
      {
        "label": "Terms of service",
        "value": "published, but our reader couldn't read it",
        "points": 7,
        "max": 10
      },
      {
        "label": "Privacy policy",
        "value": "published, but our reader couldn't read it",
        "points": 7,
        "max": 10
      }
    ],
    "unchecked": [
      "unchecked: withpersona.com marketing pages (home, pricing, security, terms of service, privacy policy, DPA and sub-processor pages) answered our reader with a bot check (HTTP 403), so the terms, the privacy policy, any SLA and any sub-processor list were not read",
      "unchecked: the trust centre at trust.withpersona.com, which needs JavaScript, so certification reports and any published sub-processors were not confirmed there",
      "unchecked: security incidents, advisories or vendor post-mortems in the last 12 months. Persona's blog and trust pages were unreadable, so the deduction is 0 on what we could read",
      "unchecked: whether customers on API versions before 2021-08-18 had notice before the backfilled `extraction-results` attribute was removed on 13 April 2026",
      "unchecked: whether the trial sign-up asks for a card (the help centre says it doesn't) and whether a failed verification counts as a billable service",
      "unchecked: sandbox rate limits, and whether the MCP server appears in the official MCP registry",
      "The versioning page names 2025-12-08 as the latest version, while the `Persona-Version` header on the create-inquiry reference page also lists 2026-09-29",
      "The API keys page describes the `mcp.access` permission as access to a Machine Learning Classification Platform, which is unrelated to the Model Context Protocol server"
    ],
    "weaknesses": [
      "Production access needs Persona's approval and a 12-month plan from $250 a month, and API-only integration is limited to Enterprise plans",
      "Eleven incidents on the status page since 10 July 2026, five marked partial outage, most on document and government ID verifications",
      "The MCP server lists all 190 tools whatever the key allows, with no annotations, toolsets or read-only subset, and OAuth isn't available yet",
      "No official server-side SDK was found. The npm, iOS, Android and React Native packages are clients for the end-user flow",
      "The marketing site, terms, privacy policy and pricing page answered our reader with a bot check, and no SLA was found in the docs or help centre"
    ],
    "agentNotes": [
      "Ask for a dedicated key limited to the permissions the task needs. A new key has every standard permission until it's limited",
      "Send `Persona-Version` on every call and an `Idempotency-Key` on every POST. A replayed key returns the first result, including a 500",
      "Treat redact calls as irreversible. They permanently delete personal data and cascade to downstream objects",
      "Stay under 300 requests a minute in production, read `RateLimit-Remaining` and `Quota-Remaining`, and back off on 429 because rejected requests still count",
      "Use `fields` and `include` to cut response size. Through MCP these parameters aren't exposed, and only 14 tools take `filter`"
    ],
    "recheck": "https://www.anchorterminal.com/builders/#disputes"
  },
  "meta": {
    "attribution": "Anchor Terminal (https://www.anchorterminal.com)",
    "docs": "https://www.anchorterminal.com/docs/",
    "generatedAt": "2026-10-08",
    "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"
  }
}