For AI agents: a documentation index is available at https://docs.coverbase.com/llms.txt. This page is also available in markdown by appending .md to the URL.
A vendor (cbvndr_...) is a parent supplier entity. A service (cbsvc_...) is a child entity under a vendor. This API creates and reads both.
All endpoints are org-scoped to the API key. See API conventions for authentication, IDs, timestamps, idempotency, and the error envelope.
To upload, version, and search bills of materials (SBOM, AIBOM, HBOM, SaaSBOM) for a vendor or service, see the Bill of Materials API.
To upload evidence documents (SOC 2 reports, policies, contracts, questionnaires) for a vendor and link them to its services and assessments, see the Documents API.
The primary contact object
Both vendor and service endpoints accept/return aprimary_contact. On a vendor it is the address the inherent-risk-questionnaire portal invite (and other vendor emails) are sent to; on a service it is the service-level point-of-contact.
On read,
primary_contact is derived from the stored contact personnel (first entry with an email); for vendors it falls back to the vendor’s security contact email. It is null if no contact is set.
Assigning people (analysts, owners, watchers)
Vendors and services accept optional personnel assignments as arrays of user IDs (cbuser_...). All are optional; on create, omit them to assign no one. On update (PATCH), supplying an array replaces the existing set; pass [] to clear it, or omit the field to leave it unchanged.
Create a vendor
POST
POST /v1/vendors201 Created with the full vendor object.
Request body
string
required
Vendor name.
string
Vendor website URL.
string
Vendor description.
string
Headquarters location.
string
How the vendor is used within the organization.
string
Your system’s identifier for this vendor (e.g. an Ariba request ID). Stored and echoed back; useful for correlation.
string[]
Risk analyst user IDs (
cbuser_...) to assign. See assigning people.string[]
Relationship owner user IDs (
cbuser_...) to assign.string[]
Watcher user IDs (
cbuser_...) to assign.object
The vendor point-of-contact (see the primary contact object). If supplied,
email is required and validated.Headers
string
required
Bearer ak_...string
Optional. Replaying the same key within 24 hours returns the original
201 body. See Idempotency.Example request
Example response
201 Created: the full vendor object (see response fields below):
Error responses
Create a child service
POST
POST /v1/vendors/{vendor_id}/services201 Created.
This endpoint does not honor
Idempotency-Key; sending the header is harmless but has no effect. Retrying after a successful create makes a second service.Path parameters
string
required
The parent vendor ID (
cbvndr_...).Request body
string
required
Service name.
string
Service description.
string
Your system’s identifier for this service.
string[]
Tag IDs (
cbtag_...) to attach to the service.string[]
Relationship owner user IDs (
cbuser_...) to assign. See assigning people.string[]
Watcher user IDs (
cbuser_...) to assign.object
The service point-of-contact (see the primary contact object). If supplied,
email is required and validated.Services do not have risk analysts; risk analysts are assigned on the parent vendor. Sending
risk_analyst_ids to a service endpoint has no effect.Example request
cURL
Example response
201 Created (see the service object):
Service object
string
Service ID (
cbsvc_...).string
The parent vendor ID.
string
Service name.
string | null
Service description.
string | null
Your external identifier, if supplied.
string[]
Tag value strings (empty array if none).
string[]
Relationship owner user IDs (
cbuser_...). Empty array if none.string[]
Watcher user IDs (
cbuser_...). Empty array if none.object | null
The service point-of-contact (see the primary contact object), or
null.object[]
Values your organization has set on this service. Empty array if none are set, and always empty on a create response. See the Custom Fields API.
integer
Unix timestamp (seconds).
Error responses
Update a service
PATCH
PATCH /v1/vendors/{vendor_id}/services/{service_id}Path parameters
string
required
The parent vendor ID (
cbvndr_...).string
required
The service ID (
cbsvc_...). Must belong to vendor_id.Request body
string
Service name.
string
Service description.
string[]
Replace the service’s tag set (tag IDs
cbtag_...).string[]
Replace the relationship owner set (user IDs
cbuser_...). Pass [] to clear.string[]
Replace the watcher set (user IDs
cbuser_...). Pass [] to clear.object
Replace the service point-of-contact (see the primary contact object). When supplied,
email is required and validated.external_id is set only at creation and is not updatable here. Services do not have risk analysts.Example request
cURL
Error responses
List vendors
GET
GET /v1/vendorsList rows are a lightweight summary; they omit the nested
services and
the personnel-assignment arrays. Fetch GET /v1/vendors/{vendor_id}
for the full object.Query parameters
string
Case-insensitive substring match on the vendor (or service) name.
string
Exact match on the vendor’s
external_id; handy for mapping your system’s
identifier back to a Coverbase vendor.integer
default:"50"
Page size, 1-200.
integer
default:"0"
Number of rows to skip, for pagination.
boolean
default:"false"
Populate
custom_fields on each row. Off by default so listing stays cheap;
single-vendor reads always include them. See the
Custom Fields API.Example request
cURL
Example response
200 OK:
object[]
The page of vendor summaries. Each carries
id, external_id, name,
website, description, hq_location, status, status_id, tags,
use_case, tier, primary_contact, custom_fields, created_at, and
updated_at. (No services or personnel arrays; see the note above.)
custom_fields is empty unless you pass include_custom_fields=true.integer
Total vendors matching the filters (across all pages).
integer
The page size used.
integer
The offset used.
Retrieve a vendor
GET
GET /v1/vendors/{vendor_id}Path parameters
string
required
The vendor ID (
cbvndr_...).Example request
cURL
Example response
Vendor object
string
Vendor ID (
cbvndr_...).string | null
Your external identifier, if set.
string
Vendor name.
string | null
Vendor website URL.
string | null
Vendor description.
string | null
Headquarters location.
string | null
Current status as a plain label string (e.g.
"Active"). Org-customizable. null if unset.string | null
ID (
cbst_...) of the current status record.string[]
Tag value strings (empty array if none).
string[]
Assigned risk analyst user IDs (
cbuser_...).string[]
Relationship owner user IDs (
cbuser_...).string[]
Watcher user IDs (
cbuser_...).string | null
Scale level ID (
cbsclvl_...) for inherent risk.string | null
Scale level ID (
cbsclvl_...) for residual risk.string | null
Type of engagement with this vendor.
string | null
How the vendor is used.
string | null
Derived vendor tier (e.g.
tier_1, tier_2). Set by workflow tiering actions; null until derived.object | null
The vendor point-of-contact (see the primary contact object), or
null.object[]
The vendor’s child services, each a service object. Empty array if none.
object[]
Values your organization has set on this vendor. Empty array if none are set. See the Custom Fields API for the value shapes. Nested services carry their own
custom_fields.integer
Unix timestamp (seconds) of creation.
integer
Unix timestamp (seconds) of last update.
Error responses
Update a vendor
PATCH
PATCH /v1/vendors/{vendor_id}Every body field is optional. Sending
null for a field is treated the same as omitting it (no change); there is no clear-to-null behavior.Request body
string
Vendor name.
string
Vendor website URL.
string
Vendor description.
string
Headquarters location.
string
ID (
cbst_...) of a status record to transition to.string[]
Replace the entire tag set, using tag IDs (
cbtag_...).string[]
Replace the risk analyst set (user IDs
cbuser_...). Pass [] to clear.string[]
Replace the relationship owner set (user IDs
cbuser_...). Pass [] to clear.string[]
Replace the watcher set (user IDs
cbuser_...). Pass [] to clear.string
Scale level ID (
cbsclvl_...) for inherent risk.string
Scale level ID (
cbsclvl_...) for residual risk.string
How the vendor is used.
string
Type of engagement.
object
Replace the vendor point-of-contact (see the primary contact object). When supplied,
email is required and validated.external_id and tier are not updatable through this endpoint (external_id is set only at creation; tier is derived by workflows).Example request
cURL
Error responses
The vendor fact sheet
The fact sheet is AI-generated company intelligence about a vendor: company profile, products, public perception, financials, leadership, compliance/legal posture, and security posture. Coverbase assembles it from public sources; in the dashboard it powers the vendor “Fact Sheet” tab and its PDF export. These two endpoints let you read that data and trigger a regeneration programmatically. Generation is asynchronous. Trigger it with the refresh endpoint, then poll the fact sheet endpoint untilstatus is "completed".
Retrieve a vendor’s fact sheet
GET
GET /v1/vendors/{vendor_id}/fact-sheet200 with status: null and every section null; it is not a 404.
Path parameters
string
required
The vendor ID (
cbvndr_...).Example request
cURL
Example response
200 OK:
Fact sheet object
string
The vendor ID (
cbvndr_...).string | null
ID of the underlying entity that holds the fact sheet data.
null if the vendor is not yet linked to an entity.string | null
string[]
Section names completed so far in the in-progress / most recent run. Empty array if none.
integer | null
Unix timestamp (seconds) the fact sheet was last updated.
null if never generated.object | null
Company profile (name, website, industry, HQ, founding year, headcount, revenue, etc.). Includes
registry_validation, the validated corporate identity, with a citation per contributing register. null until generated.object | null
Known products and offerings.
null until generated.object | null
News mentions, third-party reviews (BBB, Trustpilot, CFPB), and competitors.
null until generated.object | null
Public/private status, market cap, ticker/exchange, funding, and transactions. Includes
financial_health (the Coverbase Financial Health Score) when your organization has that feature enabled. null until generated.object | null
Leadership and headcount estimate.
null until generated.object | null
Legal actions, enforcement actions, and data-privacy certifications.
null until generated.object | null
Trust center URL, security contact, and known security incidents. Includes
external_intelligence (the measured outside-in security rating) when your organization has that feature enabled. null until generated.Section objects are best-effort and their internal shape may evolve as Coverbase improves enrichment. Treat the presence of each field defensively; any individual section can be
null even when status is "completed" if no data was found for it.Vendor intelligence sections
Three sections carry outside-in intelligence measured from registries of record and from the vendor’s own infrastructure, rather than produced by a language model. All three cite their sources. See Vendor Intelligence, Financial Health Score, and Security Intelligence for what each field means.The Financial Health Score and the measured security rating are optional features in beta, enabled per organization. Write your integration to treat
financials.financial_health and security_posture.external_intelligence as optional: both are null for an organization that does not have them enabled, and a section can be null even when status is "completed".object | null
Validated corporate identity from GLEIF, OpenCorporates, and SEC EDGAR:
official_name, registration_status, jurisdiction, incorporation_date, addresses, and the identifiers lei, duns, company_number, and cik. is_validated is true only when a register matched with high confidence; match_decision and match_confidence expose the match quality. sources carries one entry per contributing register with its identifier and a url to the record.Present for every organization. null until a lookup has run.object | null
The Coverbase Financial Health Score.
scorecard holds score (0-100), band (very_high through very_low), confidence_tier (A, B, or C), as_of, the pillars breakdown, vitals, peer_group and peer_percentile, events, overrides_applied, provenance (source → as-of date), and unavailable_sources. history carries prior scores for trend.Returned only when your organization has the Financial Health Score enabled; null otherwise.object[]
The individual inputs behind a pillar score, ordered by weighted shortfall so the input that cost the most comes first. Each entry carries:
label_key: the feature identifier (for exampleoperating_margin,registration_status), or the adverse-event typescore: what this input scored, 0-100weight: how heavily it counted within the pillarvalueandunit: the measured figure and its unit (percent,ratio,months,years,usd,days_overdue, orcategorical)detail_key: the category behind a non-numeric input (unknown,dissolved,down). When this is present,valueis not a measurement. Ascoreof 50 withdetail_key: "unknown"means the input could not be determined, not that it was measured as mediocre, so render the two cases differentlypeer_ranked:falsemeans the score came from an absolute curve rather than a peer percentile. This is the single most important caveat on any individual numberas_of: the date this input’s value was true, which is not the date of the refresh
label_key and detail_key are open sets: new identifiers appear as sources are added, so map unknown keys to a readable fallback rather than failing.object | null
The measured outside-in security rating, assembled by querying the vendor’s own public infrastructure. Top-level fields:
domain (the domain that was rated), generated_at (unix timestamp), overall_score (0-100), overall_grade (A through F), predictive_index (0-100 likelihood of imminent exploitation), factor_scores, and the raw evidence blocks: vulnerability_findings, exposed_services, breach_exposures, email_authentication, web_headers, tls_certificate, dns_health, web_reputation.Returned only when your organization has the measured security rating enabled; null otherwise.object[]
One entry per scored factor.
factor is one of patching_cadence, network_security, email_security, tls_configuration, web_application_headers, dns_health, leaked_information, ip_reputation, application_security, web_reputation, domain_squatting. score is 0-100, grade is A through F, and weight is the absolute weight the factor carried in overall_score.data_available is the field to branch on, and the one most likely to bite an integration. When it is false, the factor could not be measured on this run and was excluded from overall_score, but it still carries score: 100, grade: "A", and its normal weight, because a factor that took no deductions never left its starting value. Read data_available first and treat false as “no data”; rendering or alerting on score without that check turns an unmeasured factor into a perfect one.Only factors that were actually scored appear in the array. An organization without the measured security rating enabled gets the seven original factors and no entries for email_security, tls_configuration, web_application_headers, or domain_squatting at all.findings carries the structured evidence (below); drivers carries the same evidence pre-rendered as English sentences, intended for prompt context rather than display.object[]
The observed facts behind a factor’s score, structured so they can be localized:
code: the finding identifier, which selects the sentence (dmarc_policy_none,dkim_weak_key,hsts_missing,cookie_not_secure, …)severity:critical,high,moderate,low, orinfosubject: what the finding is about, such as a host, a DKIM selector, a cookie name, or a CVEdetail: the observed value, rendered verbatim in every locale (p=none,1024-bit,TLSv1.0)penalty: points this finding cost its factorevidence_url: a link to the record, where one existsfirst_seen/last_seen: ISO-8601 dates carried across refreshes.first_seenis how long the finding has stood open, which is often more informative than the finding itself
code is an open set and grows as probes are added. Map an unrecognized code to subject and detail, which are always human-readable, rather than dropping the finding.Enabling the measured security rating changes
overall_score for existing vendors: the rating goes from seven factors to eleven, with new weights. If you store or alert on overall_score, expect a one-time shift when the feature is turned on for your organization.Error responses
Refresh a vendor’s fact sheet
POST
POST /v1/vendors/{vendor_id}/fact-sheet/refresh202 Accepted and status: "pending". Existing fact sheet data is cleared at the start of the run, so a GET made right after a refresh reflects the pending state. Poll the fact sheet endpoint until status is "completed".
This endpoint is not idempotent; each call queues a fresh generation run.
Path parameters
string
required
The vendor ID (
cbvndr_...).Example request
cURL
Example response
202 Accepted:
string
The vendor ID (
cbvndr_...).string | null
ID of the underlying entity, or
null when the vendor is being linked to one for the first time as part of this run.string
Always
"pending" on a successful enqueue.string
Human-readable acknowledgement.