Skip to main content
Glama

Open Agent Guide

Search Products

search_products
Read-onlyIdempotent

Search AI-native products by free text, membership (category, job-to-be-done, capability, publisher organization or person, SDK language, license, certification, access control, agent skill, benchmark, documentation resource, MFA type), AI-native tier, active-offer billing_model/payment_method, or affordance booleans (has_mcp selects hosted MCP servers; has_mcp_url narrows that to a dialable endpoint url; has_api_spec and spec_type select API specs). Membership and reference filters take one slug or an array (OR within a filter; filters combine with AND). view="compact" returns a curated shortlist projection for a first pass (identity, the rating and coverage signals, the affordance booleans, and the connect summary) and cannot combine with fields; fields trims the payload to an explicit set and expand resolves relation *_ids into inline objects. Returns a cursor-paginated page whose products carry ai_native_score (0-100), ai_native_tier (S/A/B/C/D), the research-coverage axis coverage_score (0-100) and derived unrated boolean, the attested_absences list, and the read booleans has_mcp, has_mcp_url, has_sdk, has_api_spec, has_cli, and has_a2a (each reports catalog record presence), and free_plan (True when the product has an active free-tier offer; null when none is recorded, so rank a shortlist by cost without opening each product — the value is never false because offers carry no attested absence). Evidence-based filters exclude unknown values in both directions. Negative relation filters require whole-relation absence evidence. Negative offer filters return zero matches until offer absence is supported. Read presence booleans alone do not establish filter evidence. The page carries filter_breakdown: matched, recorded_nonmatch, and unknown counts for each supported filter. Previously covered positive requests also retain recorded_negative as an alias of recorded_nonmatch. Sum only the three canonical counts. Refinements include spec_type, sdk_language, has_mcp_url, every data_export set, and certification, access-control, or agent-skill membership. A different relation member leaves the requested value unknown. get_product has the full reading guidance. The REST *_id UUID filters and the web-only following filter are intentionally not exposed here; use the HTTP API for those.

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": "search_products"
  }
}

Input Schema

NameRequiredDescriptionDefault
tierNoAI-native tier: S, A, B, C, D, or Unrated. A letter matches only products whose coverage clears the research threshold; Unrated matches the rest. Accepts one slug or a nonempty array of slugs; an array matches any (OR). Separate filters combine with AND.
viewNoOpt-in named projection, mirroring the REST `view` parameter. `compact` returns the shortlist field set for a first pass: product identity, the ai_native_tier, ai_native_score, coverage_score, and unrated rating signals, the has_* affordance booleans, and the connect summary. It omits the *_ids arrays and the null-heavy decision scalars. Pagination, the weak-match signal, and the suggested categories and jobs are unchanged. Cannot combine with fields. Omit for the full record.
queryNoDeprecated alias for search; renames to the same catalog parameter. Use search instead (ADR-0033).
cursorNoOpaque cursor from a previous page's response.
expandNoComma-separated product relations to expand inline as full objects rather than opaque *_ids, following the REST ?expand= grammar. Example: mcp_servers,developer_kits,capabilities. Lets you resolve a result's affordance IDs (SDK languages, spec URLs, MCP endpoints) without a follow-up get_product call. An invalid value returns the allowed set. Omit it for the lighter id-only page.
fieldsNoComma-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.
pay_goNoAccepts true/false, 1/0 (JSON boolean or number), yes/no, or on/off, ignoring case and surrounding whitespace. Empty values are invalid. Omit to leave unfiltered.
personNoFilter by publisher person slug. Accepts one slug or a nonempty array of slugs; an array matches any (OR). Separate filters combine with AND.
searchNoFree-text search across products. Search matches product name, description, and related capability and job-to-be-done names by meaning, so query the capability a product provides, such as "weather forecast", not the user's whole goal, such as "plan a family outing based on weather". When set, the response carries a page-level `strong_match` boolean: `false` means every returned product surfaced only through approximate meaning-similarity, or there were none, so narrow the search before relying on a result. On a `false` match the response also lists `suggested_categories` and `suggested_jobs_to_be_done` (lexical matches most-populated first, semantic matches nearest-intent first; each with a `product_count` and a `match_kind`) — pass a suggestion's slug to the matching filter to switch from free-text search to exact membership. `match_kind` is `lexical` when the term's label contained your wording and `semantic` when it is a close match on intent worth checking. An empty suggestion list means neither your wording nor its intent matched a term; retry with a shorter query or browse the taxonomy.
has_a2aNoAccepts true/false, 1/0 (JSON boolean or number), yes/no, or on/off, ignoring case and surrounding whitespace. Empty values are invalid. Omit to leave unfiltered.
has_cliNoAccepts true/false, 1/0 (JSON boolean or number), yes/no, or on/off, ignoring case and surrounding whitespace. Empty values are invalid. Omit to leave unfiltered.
has_mcpNoAccepts true/false, 1/0 (JSON boolean or number), yes/no, or on/off, ignoring case and surrounding whitespace. Empty values are invalid. Omit to leave unfiltered.
has_sdkNoAccepts true/false, 1/0 (JSON boolean or number), yes/no, or on/off, ignoring case and surrounding whitespace. Empty values are invalid. Omit to leave unfiltered.
licenseNoFilter by license slug, e.g. apache-2-0. Accepts one slug or a nonempty array of slugs; an array matches any (OR). Separate filters combine with AND.
sandboxNoAccepts true/false, 1/0 (JSON boolean or number), yes/no, or on/off, ignoring case and surrounding whitespace. Empty values are invalid. Omit to leave unfiltered.
web_mcpNoAccepts true/false, 1/0 (JSON boolean or number), yes/no, or on/off, ignoring case and surrounding whitespace. Empty values are invalid. Omit to leave unfiltered.
categoryNoFilter by category slug. Accepts one slug or a nonempty array of slugs; an array matches any (OR). Separate filters combine with AND.
mfa_typeNoFilter by multi-factor-auth type slug. Accepts one slug or a nonempty array of slugs; an array matches any (OR). Separate filters combine with AND.
benchmarkNoFilter by benchmark slug. Accepts one slug or a nonempty array of slugs; an array matches any (OR). Separate filters combine with AND.
free_planNoTrue matches products with an active Offer containing a structured pricing option with price_kind: free. False returns zero matches until the catalog supports offer absence. Missing offers remain unknown. Omit to leave unfiltered. Prose-only free allowances do not match without a qualifying structured option. Offers can have quotas, eligibility restrictions, noncommercial terms, or attribution requirements. Inspect offer terms and url. Example: search_products(free_plan=true, expand="offers"). Accepts true/false, 1/0, yes/no, or on/off, ignoring case and surrounding whitespace. Empty values are invalid.
page_sizeNoResults per page (1-100).
spec_typeNoAPI specification type: openapi, grpc, or graphql. Supply an array for OR matching. Invalid values return the allowed set.
capabilityNoDeprecated alias for capability_slug. Supply only one spelling.
agent_skillNoFilter by agent-skill slug. Accepts one slug or a nonempty array of slugs; an array matches any (OR). Separate filters combine with AND.
data_exportNoRecorded data export support: full, partial, or none. Supply an array for OR matching. Unknown values are excluded.
has_mcp_urlNoAccepts true/false, 1/0 (JSON boolean or number), yes/no, or on/off, ignoring case and surrounding whitespace. Empty values are invalid. Omit to leave unfiltered.
no_trainingNoAccepts true/false, 1/0 (JSON boolean or number), yes/no, or on/off, ignoring case and surrounding whitespace. Empty values are invalid. Omit to leave unfiltered.
open_sourceNoMatches the recorded boolean; excludes unknown. Accepts true/false, 1/0, yes/no, or on/off, ignoring case and whitespace. Omit to leave unfiltered.
has_api_specNoAccepts true/false, 1/0 (JSON boolean or number), yes/no, or on/off, ignoring case and surrounding whitespace. Empty values are invalid. Omit to leave unfiltered.
organizationNoFilter by publisher organization slug. Accepts one slug or a nonempty array of slugs; an array matches any (OR). Separate filters combine with AND.
sdk_languageNoFilter by a product SDK's programming-language slug, e.g. python. Accepts one slug or a nonempty array of slugs; an array matches any (OR). Separate filters combine with AND.
billing_modelNoFilter by an active offer's billing-model slug. Accepts one slug or a nonempty array of slugs; an array matches any (OR). Separate filters combine with AND.
certificationNoFilter by compliance certification slug, e.g. soc-2. List tool: list_compliance_certifications. Accepts one slug or a nonempty array of slugs; an array matches any (OR). Separate filters combine with AND.
self_hostableNoMatches the recorded boolean; excludes unknown. Accepts true/false, 1/0, yes/no, or on/off, ignoring case and whitespace. Omit to leave unfiltered.
access_controlNoFilter by access-control slug. Accepts one slug or a nonempty array of slugs; an array matches any (OR). Separate filters combine with AND.
job_to_be_doneNoFilter by job-to-be-done slug. Accepts one slug or a nonempty array of slugs; an array matches any (OR). Separate filters combine with AND.
open_standardsNoAccepts true/false, 1/0 (JSON boolean or number), yes/no, or on/off, ignoring case and surrounding whitespace. Empty values are invalid. Omit to leave unfiltered.
payment_methodNoFilter by an active offer's payment-method slug. Accepts one slug or a nonempty array of slugs; an array matches any (OR). Separate filters combine with AND.
capability_slugNoFilter by capability slug. Accepts one slug or a nonempty array of slugs; an array matches any (OR). Separate filters combine with AND.
headless_signupNoAccepts true/false, 1/0 (JSON boolean or number), yes/no, or on/off, ignoring case and surrounding whitespace. Empty values are invalid. Omit to leave unfiltered.
zero_data_retentionNoAccepts true/false, 1/0 (JSON boolean or number), yes/no, or on/off, ignoring case and surrounding whitespace. Empty values are invalid. Omit to leave unfiltered.
documentation_resourceNoFilter by documentation-resource slug. Accepts one slug or a nonempty array of slugs; an array matches any (OR). Separate filters combine with AND.

Output Schema

NameRequiredDescriptionDefault

No arguments

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observed

TDQS

A3.9/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations already cover the safety profile (readOnly, idempotent, non-destructive), but the description adds genuinely non-obvious behavior: evidence-based filters exclude unknown values in both directions, negative relation filters require whole-relation absence evidence, negative offer filters return zero matches until offer absence is supported, and read presence booleans alone do not establish filter evidence. That is substantive disclosure an agent could not infer from the schema. Some space is spent restating return fields that the output schema already carries, which keeps it from a 5.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness3/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The search scope is front-loaded, but the body is a single dense run-on paragraph that mixes invocation guidance with return-field documentation and an opaque sentence ('Refinements include spec_type, sdk_language, has_mcp_url, every data_export set...'). Several clauses duplicate the schema's own OR/AND wording. It is information-dense rather than padded, but it is hard to scan and would benefit from structure.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a 42-parameter tool with an output schema, the description covers the decision-relevant ground: pagination via cursor, projection controls, filter-combination semantics, evidence caveats, and an explicit statement of what is deliberately not exposed. The remaining gap is that it never notes whether an empty/filtered query is valid or what an unfiltered call returns by default, which matters when all 42 parameters are optional.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, so the schema already defines every parameter, and the description largely repeats the OR-within/AND-across rule. It does add meaning the schema lacks: the evidence rules that govern what a filter will and will not match, the `free_plan` null-vs-false distinction, and the `view="compact"`/`fields` exclusivity plus the interaction between `expand` and `fields`. Those additions justify one step above the 3 baseline.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

The opening sentence states a specific verb and resource ('Search AI-native products') and enumerates the filterable dimensions (free text, membership, tier, offer billing/payment, affordance booleans). It also points to `get_product` for full reading guidance, giving a weak sibling boundary. It does not, however, contrast itself with the `list_*` taxonomy siblings, which an agent could plausibly confuse with membership filters.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

It gives real selection guidance: query the capability a product provides rather than the user's whole goal, use `view="compact"` for a first pass, and use `expand` to avoid a follow-up `get_product` call. The `strong_match` false path tells the agent to narrow before relying on results, and to convert a suggestion slug into a membership filter. It stops short of an explicit when-not-to-use-this-tool statement beyond the note that REST-only `*_id` and `following` filters are not exposed here.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Try in Browser

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources