Get Product
get_productFetch one catalog product by UUID or slug with its AI-native scores
(ai_native_score 0-100, ai_native_tier S/A/B/C/D) and related
organization, capabilities, and integrations (MCP servers, SDKs with languages
and licenses, CLIs, A2A servers, API specs, plugins) expanded inline as objects.
Also returned: a second research-coverage axis — coverage_score (0-100)
and the derived boolean unrated. Coverage tells you how thoroughly the
product was researched before you trust the grade; unrated is true when
research coverage is too low to assign a letter grade (present as "Unrated /
needs research"). Plus the always-present attested_absences list — fields a
dated check attested absent (each a field_key or field_key:label for a
relation-scoped negative), the read booleans has_mcp, has_mcp_url,
has_sdk, has_api_spec, has_cli, has_a2a (each reads back its
filter), and free_plan (True for an active free-tier offer, else null;
never false).
has_mcp_url is true when a cataloged MCP server has a dialable endpoint URL.
It is false when no MCP endpoint URL is cataloged.
documented_task_completions contains all assessed jobs. Match each
selected job to job_to_be_done_slug or job_to_be_done_id. Show its
verdict, scope, and conditions with its assessment. The verdict
includes delegation limits. No matching record means
unassessed; a matching record with null completion_support means unknown.
Neither establishes inability to complete the job. Keep other jobs secondary.
If you use fields, include documented_task_completions.
The identifier is forwarded to the canonical {id-or-slug} detail route,
which owns the missing/ambiguous/invalid errors: an unknown value
is 404, a slug matching multiple products is 409, a malformed slug is
400 — each a structured tool error with the catalog's code and
detail, consistent with get_product_coverage. fields trims the
payload; the default affordance relations are still expanded.
A missing or unknown value is JSON null on every field type — the
nullable booleans, the status enums (MFA, data-retention, sandbox,
portability), and the URL/text fields alike — so never read a blank as a
negative; the real negative is always explicit (false, or
no/none/unavailable). attested_absences then flags which
negatives are cited: listed is a cited "no", an explicit negative not listed
is recorded but uncited, and a null is unknown. The full attestation
records (method, source_url, checked_at) are not returned by
default; pass expand=absence_attestations to include them, and use
expand to resolve any other relation *_ids into inline objects.
Tool call template
This tool declares no required parameters. Add optional parameters from the input schema as needed.
{
"id": 1,
"jsonrpc": "2.0",
"method": "tools/call",
"params": {
"arguments": {},
"name": "get_product"
}
}Input Schema
| Name | Required | Description | Default |
|---|---|---|---|
| id | No | Alias for product, matching the REST {id-or-slug} route. Accepts a UUID or slug. | |
| expand | No | Comma-separated product relations to expand inline beyond the affordance and documented_task_completions relations this tool always expands, following the REST ?expand= grammar. Example: compliance_certifications,offers,absence_attestations. Your values are added to the defaults, so name only the extras. An invalid value returns the allowed set. | |
| fields | No | Comma-separated top-level response fields to include, following the REST ?fields= grammar. The response always keeps id. Narrows the payload for compact reads; an expanded relation is still returned only when its field is in the selection. Omit for the full record. | |
| product | No | The product's UUID or slug. | |
| product_id | No | Legacy UUID-only argument. Use product instead. |
Output Schema
| Name | Required | Description | Default |
|---|---|---|---|
No arguments | |||