# Company Subprocessors (`company.subprocessors`)

Finds a company's published subprocessor list from its domain (or reads the page you give) and returns each vendor with its purpose, data locations and the list's last-updated date.

Docs: https://primaryapi.com/docs/api-reference/company/subprocessors

- Endpoint: `POST https://api.primaryapi.com/v1/run/company.subprocessors`
- Category: Compliance & Records · Platform: `company` · Kind: scrape · Status: live
- Price: 1 credit per call; up to 13 credits if a call is retried elsewhere. A quote’s `maxCredits` is the most a given call can cost.
- Billing unit: per successful call
- Reliability: routed to a healthy upstream and retried automatically when an attempt fails; live 30-day status: `GET https://api.primaryapi.com/v1/status/apis?sku=company.subprocessors`
- Execution: durable: may answer 202 and finish in the background; poll `GET /v1/requests/{id}`

## Input

| Field | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| `domain` | string | no |  | max 2048 chars | Company domain such as "stripe.com". A URL or email address is reduced to its domain; "www." is dropped. |
| `url` | string | no |  | max 2048 chars; format: uri | The subprocessor page, DPA or trust-center URL when you know it. Without it the list is searched for on the domain. |

Unknown fields are rejected with `400 invalid_input`. The routing controls `preferLatencyUnderMs` and `maxCredits` may be sent in the same JSON body.

## Output

When `output.found` is `true`, `output.data` has this shape:

| Field | Type | Description |
|---|---|---|
| `domain` | string \| null | Company domain the list belongs to. |
| `sourceUrl` | string | Page the list was read from. |
| `format` | enum | html_table = HTML table(s); trust_center = a hosted trust center; document = text or PDF. |
| `trustCenterPlatform` | string \| null | Hosted trust-center platform when the page is one (safebase, secureframe, …). |
| `lastUpdated` | string \| null | The "last updated" / "effective" date exactly as the page states it, or null when it states none. |
| `lastUpdatedDate` | string \| null | The same date as ISO 8601 (YYYY-MM-DD, or YYYY-MM when the page gives only a month). |
| `notice` | object |  |
| `notice.methods` | enum[] | How the company says it announces changes: email sign-up, RSS feed, trust-center subscription, this page. |
| `notice.periodDays` | integer \| null | Advance notice of new subprocessors in days, when the page states one ("at least 30 days"). |
| `notice.text` | string \| null | The sentence that states the notice method or period (at most 300 characters). |
| `vendorCount` | integer |  |
| `vendors` | object[] | Sorted: third parties first, then affiliates; each by name. |
| `vendors[].name` | string | Vendor or legal entity name as listed. |
| `vendors[].purpose` | string \| null | What the vendor does for the company, as stated (whitespace normalised). |
| `vendors[].locations` | string[] | Countries or regions where it processes data, as stated (one entry per place). |
| `vendors[].category` | enum \| null | Inferred from the stated purpose with a fixed keyword list; null when nothing matches. |
| `vendors[].group` | string \| null | Section label the page lists the vendor under, as stated. |
| `vendors[].type` | enum | "affiliate" for the company's own group entities, else "third_party". |
| `vendors[].website` | string \| null | Vendor website when the page gives one. |
| `readWith` | enum | parser = read from the page structure; ai = read with AI because the page is not a clean table. |
| `truncated` | boolean | true when the page was too long and only part of it was read. |

When the thing you asked about does not exist (or a search matches nothing) the call still succeeds: `output.found` is `false`, `output.data` is `null` and `output.reason` says why (`not_found`, `private`, `no_results`, …). Misses are free unless the notes below say misses are billed.

## Example request

```bash
curl -sS -X POST "https://api.primaryapi.com/v1/run/company.subprocessors" \
  -H "Authorization: Bearer $PRIMARYAPI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"domain":"github.com"}'
```

## Example response

_Values are placeholders that follow the output schema; `credits` is what this call would cost at the usual price._

```json
{
  "output": {
    "found": true,
    "data": {
      "domain": "github.com",
      "sourceUrl": "https://docs.github.com/en/site-policy/privacy-policies/github-subprocessors",
      "format": "html_table",
      "trustCenterPlatform": null,
      "lastUpdated": null,
      "lastUpdatedDate": null,
      "notice": {
        "methods": [
          "webpage"
        ],
        "periodDays": 30,
        "text": "publishes the names of any new subprocessors … at least 30 days in advance"
      },
      "vendorCount": 2,
      "vendors": [
        {
          "name": "Amazon Web Services Inc (AWS)",
          "purpose": "Cloud Hosted Infrastructure, Data Hosting, AI Inference and AI Services",
          "locations": [
            "United States"
          ],
          "category": "hosting",
          "group": null,
          "type": "third_party",
          "website": null
        },
        {
          "name": "Anthropic PBC",
          "purpose": "AI Inference and AI Services",
          "locations": [
            "United States"
          ],
          "category": "ai",
          "group": null,
          "type": "third_party",
          "website": null
        }
      ],
      "readWith": "parser",
      "truncated": false
    }
  },
  "provider": "PrimaryAPI",
  "credits": 1,
  "items": 1,
  "requestId": "req_01JBXK6Y2N9V8T4R3Q2P1M0A7Z"
}
```

## Pricing and reliability

- Usual price: 1 credit per call
- Price ceiling: up to 13 credits if a call is retried elsewhere. A quote’s `maxCredits` is the most a given call can cost; the `maxCredits` routing control caps it per call
- Live 30-day status, success rate, uptime and latency: `GET https://api.primaryapi.com/v1/status/apis?sku=company.subprocessors`

## Errors to handle

`invalid_input`, `unauthorized`, `insufficient_balance`, `key_limit_reached`, `trial_cap_reached`, `forbidden`, `not_found`, `idempotency_conflict`, `rate_limited`, `all_sources_failed`, `service_unavailable`, `timeout`. See https://primaryapi.com/docs/errors.

## Notes

No list found on the site is a free miss (reason no_subprocessor_list); a page that only renders in a browser is a free miss (reason page_not_readable). Lists published as HTML tables or on common hosted trust centers cost 1 credit; other layouts (lists, prose, PDF) are read with AI and cost more. Text read with AI is checked against the page: names and quotes that do not appear in it are dropped. Treat extracted facts as a starting point and confirm them in the source before relying on them. Pages are read the way a polite crawler would: robots.txt is honoured (a disallowed page is a free miss with reason robots_disallowed), requests are paced per site, and only facts, short quotes and source links are returned. Outputs are stable between runs of an unchanged page (sorted lists, normalised whitespace, no timestamps), so the API can be scheduled as a monitor that fires only when the content changes.

