Skip to main content
Glama

Search the apinoa docs

search_docs
Read-only

Find how to perform tasks by looking up matching operations, parameters, per-call costs, and response fields before writing code; no API calls are made.

Instructions

Find how to do a task with apinoa: the matching operations as you call them in execute (e.g. apinoa.tiktok.posts({ hashtag })), every parameter, the cost per call, and the fields of what they return. Free: it calls nothing. Use it before writing code. Pass words ('tiktok hashtag videos', 'instagram reels of an account', 'amazon product reviews'), a platform, or a type name ('Post'); with neither you get the list of every operation.

Platforms: aliexpress, amazon, instagram, linkedin, reddit, shopify, threads, tiktok, tiktokshop, walmart, x, youtube.

Input Schema

NameRequiredDescriptionDefault
queryNoWhat you want to do or read, in a few words, or a type name.
platformNoOnly this platform's operations.

Schema Changelog

Changes observed during successful MCP inspections.

  1. Addedv0.2.1

TDQS

A4.5/5.0
Behavior4/5

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

Annotations already declare readOnlyHint=true and openWorldHint=false; the description adds the meaningful behavioral fact that it performs no calls and therefore costs nothing, plus what a no-argument query returns. It does not cover result ordering, truncation, or result-set size limits, so it stops short of 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.

Conciseness4/5

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

Front-loaded with the payload summary, then usage guidance, then input forms, then the platform list — a sensible order with little waste. The first sentence is dense and runs long, packing the return contents into one clause, which slightly hurts scannability.

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?

No output schema exists, and the description compensates by describing what a result contains (operations, signatures, parameters, cost per call, response fields). It is nearly complete for a read-only search tool; only result formatting and limits are unaddressed.

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 baseline is 3, but the description adds real semantics: example query strings, the notion of a type name, the enumerated platform values, and — crucially — the zero-argument behavior (return everything), which the schema does not convey. The platform list is the main value-add over the generic schema text.

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

Purpose5/5

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

States a specific verb and resource (find operations/parameters/costs/return fields in the apinoa docs) and explicitly frames itself as the documentation lookup paired with the sibling execute tool ("the matching operations as you call them in execute"). An agent can distinguish it from execute and translate_image without opening any schema.

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

Usage Guidelines5/5

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

Gives an explicit directive ("Use it before writing code"), the condition that makes it safe/cheap ("Free: it calls nothing"), and concrete input strategies: keywords, a platform, a type name, or nothing at all to list every operation. Effectively names the alternative path (execute) and when this precedes it.

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

Deploy Server

Other Tools