mcp-retrieval
Qué es
mcp-retrieval es un servidor Model Context Protocol escrito en Go. Expone capacidades de recuperación web a cualquier cliente compatible con MCP (Claude Desktop, agentes de IDE, aplicaciones LLM personalizadas) como tres herramientas de solo lectura. Internamente utiliza la librería retrieval-go para buscar en la web y obtener páginas, devolviendo resultados como Markdown limpio listo para entregar a un modelo.
La librería no necesita claves API: la búsqueda web pasa por DuckDuckGo Lite, la búsqueda de imágenes por Bing Images, y la obtención de páginas ejecuta el HTML a través de un extractor de legibilidad antes de convertirlo a Markdown. Para mantenerse fiable contra la protección de bots, suplanta a navegadores reales a nivel de TLS y puede rotar tanto huellas de navegador como proxies — ver Motor de recuperación.
Ambos transportes que soporta el SDK de MCP están disponibles y exponen el mismo conjunto de herramientas:
stdio — el cliente lanza el binario y habla por stdin/stdout (el predeterminado, ideal para clientes de escritorio).
http — un servidor HTTP transmisible de larga duración (útil para despliegues remotos/compartidos).
Related MCP server: mcp-web-calc
Herramientas
Herramienta | Descripción |
| Ejecuta una o más consultas en paralelo y devuelve fragmentos deduplicados y reordenados por consulta con enlaces. |
| Ejecuta una o más consultas de imágenes en paralelo y devuelve resultados de imágenes deduplicados por consulta. |
| Descarga una o más páginas en paralelo y devuelve el texto principal del artículo como Markdown. |
Las tres están anotadas como de solo lectura. Cada herramienta devuelve una carga JSON estructurada que coincide con su esquema de salida; el SDK refleja el mismo JSON en el bloque de contenido de texto para clientes que no leen structuredContent.
web_search
Parámetro | Tipo | Predeterminado | Notas |
|
| — | Obligatorio. Se ejecutan en paralelo. |
|
|
| Fragmentos por consulta, limitado a |
|
|
| Tiempo de espera de toda la llamada; limitado a |
|
| — | Filtro de frescura: |
web_search_images
Parámetro | Tipo | Predeterminado | Notas |
|
| — | Obligatorio. Se ejecutan en paralelo. |
|
|
| Imágenes por consulta, limitado a |
|
|
| Tiempo de espera de toda la llamada; limitado a |
|
| — | Filtro de frescura: |
web_scrape
Parámetro | Tipo | Predeterminado | Notas |
|
| — | Obligatorio. Se descargan en paralelo. |
|
|
| Respetar el |
|
|
| Tiempo de espera de toda la llamada; limitado a |
|
|
| Eliminar enlaces Markdown del texto. |
|
|
| Truncar el texto de la página a N caracteres, limitado a |
Tanto las listas
queries/urlsestán limitadas amax_queries(10) elementos por llamada. Las consultas deben tener ≤ 512 caracteres; las URLs ≤ 2048 caracteres y solohttp/https.
Resultados y recuentos
Cada llamada se distribuye en la lista de entrada y devuelve una entrada por consulta/URL, cada una con su propio status — success, failed o timeout — de modo que un fallo parcial aún devuelve los elementos que sí funcionaron.
count es el número de elementos realmente devueltos, y puede ser menor que el max_results / max_images solicitado: los duplicados dentro de los resultados de una sola consulta se eliminan antes de aplicar el límite, y el proveedor puede simplemente tener menos elementos para dar. Un count más pequeño es un resultado normal, no un error.
La deduplicación es por consulta, no entre consultas. Cada entrada se deduplica por sí misma, por lo que un enlace encontrado por dos de las consultas en la misma llamada aparece en ambas entradas — deduplica la unión tú mismo si lo necesitas.
Errores
Los fallos a nivel de solicitud se devuelven como resultado de herramienta con isError: true y un mensaje de texto plano, no como un error JSON-RPC — el modelo lee el mensaje y puede corregir la llamada por sí mismo. Los fallos por elemento nunca hacen esto; permanecen dentro de la carga como status: "failed" / "timeout".
Una llamada falla por completo solo cuando la entrada se rechaza antes de que comience cualquier trabajo, o cuando todos los elementos de la misma fallan:
Mensaje | Significado |
| Los argumentos no pasaron la validación. |
| La lista supera |
| Una consulta vacía, o una lista |
| Una consulta supera los 512 caracteres. |
| Una URL está malformada, supera los 2048 caracteres, o no es |
|
|
| El proveedor respondió con un código de estado inesperado. |
| Todas las URLs fallaron. Las causas individuales se registran en |
| Todas las consultas fallaron. |
| Cualquier cosa no clasificada. |
Los mensajes de fallo total deliberadamente no distinguen los tiempos de espera de otras causas: un lote mixto puede fallar por varias razones a la vez, y el status por elemento ya lleva ese detalle siempre que al menos un elemento sobreviva.
Limitaciones conocidas
web_scrapesolo maneja HTML. Las páginas pasan por un extractor de legibilidad, que necesita marcado de artículo, por lo que las respuestastext/plainno producen nada y vuelven comostatus: "failed". Los hosts de archivos crudos son el caso común:raw.githubusercontent.com,github.com/.../raw/...,cdn.jsdelivr.net. Extrae la página renderizada en lugar del archivo crudo.La relevancia de
web_search_imagesno está garantizada. Para algunas consultas, Bing Images sirve una página que no es un conjunto de resultados, y se analiza como si lo fuera — la herramienta entonces devuelve imágenes no relacionadas constatus: "success". Trata los resultados de imágenes como de mejor esfuerzo y verifícalos antes de mostrarlos a un usuario.Sin JavaScript. Las páginas se obtienen tal cual; el contenido renderizado en el lado del cliente es invisible para el extractor.
Inicio rápido
Instalación
Elige la que mejor se adapte — todas dan el mismo servidor.
Contenedor (sin necesidad de toolchain de Go):
docker pull ghcr.io/role1776/mcp-retrieval:latestBinario precompilado — descarga el archivo para tu plataforma desde la última versión, descomprímelo y pon mcp-retrieval en tu PATH.
Paquete MCP — para clientes que instalan archivos .mcpb, descarga mcp-retrieval_<version>_<os>_<arch>.mcpb desde la última versión y ábrelo con tu cliente. El paquete lleva el binario compilado, por lo que no necesita ni Docker ni Go. Elige el archivo que coincida con tu sistema operativo y arquitectura de CPU: un paquete contiene un binario nativo.
Desde el código fuente:
go install github.com/Role1776/mcp-retrieval/app/cmd/mcp-retrieval@latest # needs Go 1.25.5+O construye el binario en el lugar (el módulo Go vive en app/):
make build # -> bin/mcp-retrievalEjecutar
# defaults: stdio transport, no configuration needed
./bin/mcp-retrieval
# with an explicit env file
./bin/mcp-retrieval -env /absolute/path/to/.envLa única bandera es opcional:
Bandera | Significado |
| Ruta a un archivo |
Conectar un cliente MCP (stdio)
Apunta tu cliente al binario compilado. Ejemplo de configuración de Claude Desktop:
{
"mcpServers": {
"retrieval": {
"command": "/absolute/path/to/mcp-retrieval",
"env": {
"MAX_RESULTS": "20"
}
}
}
}El bloque env es opcional: "command" por sí solo es suficiente.
Conexión de un cliente MCP (contenedor)
Ejecuta la imagen en stdio. La configuración sigue viajando a través del bloque env, pero Docker necesita que cada variable se nombre en la línea de comandos con -e para que llegue al proceso:
{
"mcpServers": {
"retrieval": {
"command": "docker",
"args": [
"run", "-i", "--rm",
"-e", "MAX_RESULTS",
"-e", "DEFAULT_TIMEOUT_MS",
"ghcr.io/role1776/mcp-retrieval:latest"
],
"env": {
"MAX_RESULTS": "20",
"DEFAULT_TIMEOUT_MS": "5000"
}
}
}
}-i es obligatorio: sin él, el contenedor no recibe stdin y el cliente ve morir al servidor de inmediato. Los clientes que se instalan desde el MCP Registry construyen esta invocación por sí mismos y solicitan las variables declaradas en server.json.
Ejecución sobre HTTP
Establece MCP_TRANSPORT=http y el servidor escucha en SERVER_PORT en MCP_PATH (por defecto http://localhost:8080/mcp).
Configuración
Todo se configura mediante variables de entorno, y cada valor se valida antes del arranque: un valor no numérico o no positivo es un error de arranque. Las relaciones entre límites no se comprueban en el arranque; consulta Límites. Las variables ya presentes en el entorno tienen prioridad sobre un archivo .env, por lo que el bloque env de un cliente MCP siempre tiene efecto. Cada campo tiene un valor predeterminado sensato, de modo que el servidor se ejecuta sin configuración alguna (transporte stdio).
Consulta .env.example para ver la lista completa con sus valores predeterminados, lista para copiar a .env.
Servidor MCP
Env | Default | Notas |
|
|
|
|
| Nombre del servidor anunciado a los clientes. |
|
| Ruta HTTP (solo transporte http). |
La versión anunciada a los clientes no es configurable: se graba en el binario en tiempo de compilación a partir de la etiqueta git.
Servidor HTTP (solo transporte http)
Env | Default |
|
|
|
|
|
|
Cliente HTTP y proxy
Env | Default | Notas |
|
| Agrupación de conexiones HTTP. |
| — | Opcional. Si se establece, las solicitudes se enrutan a través de un proxy de sesión rotatoria. |
| — | Obligatorio cuando se establece |
| — | Obligatorio cuando se establece |
| — | Obligatorio cuando se establece |
| — | Obligatorio cuando se establece |
Cuando se configura un proxy, cada solicitud saliente recibe un identificador de sesión único añadido al inicio de sesión, de modo que el proveedor upstream rota la IP de salida por solicitud.
Límites
Env | Default |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Cada valor se comprueba por separado: debe ser mayor que cero, pero los tríos DEFAULT_*, MIN_* y MAX_* no se verifican entre sí en el arranque. Un conjunto incoherente no detiene el servidor; se reconcilia por solicitud en su lugar:
un valor que el llamante omite, o pasa como cero o negativo, recurre al
DEFAULT_*correspondiente;el resultado se limita luego a
[MIN_*, MAX_*], de modo que unDEFAULT_*mayor que suMAX_*simplemente produceMAX_*;si
MIN_*supera aMAX_*, gana el máximo.
Por lo tanto, el límite efectivo siempre está dentro del máximo configurado, y una mala configuración degrada a un servidor funcional en lugar de un arranque fallido. La contrapartida es que degrada silenciosamente: un error tipográfico como MAX_RESULTS=2 en lugar de 20 no produce ninguna advertencia, solo respuestas discretamente más pequeñas. Vale la pena revisar estos valores cuando los resultados parecen truncados.
Registro
Env | Default | Notas |
|
|
|
Arquitectura
El proyecto sigue una estructura limpia y en capas. Las dependencias apuntan hacia adentro, hacia el dominio, y cada capa se comunica con la siguiente a través de interfaces.
app/ the Go module: sources plus its build files
(Dockerfile, .dockerignore, .goreleaser.yaml)
cmd/mcp-retrieval/main.go entry point: parse flags, load config, run app
internal/
app/ wiring + lifecycle (build server, run, graceful shutdown)
config/ config loading (.env → env vars → validate)
domain/ core types (Query, Link, Document, Snippet, Image) and errors
dto/web/ request/response shapes for the MCP tools
transport/mcp/ MCP layer
router/ registers every tool group on the MCP server
web/ tool handlers
utils/ schema helpers and error → tool-result mapping
usecase/web/ business logic: validation, parallelism, timeouts, dedupe/limit/rerank
adapter/web/ retrieval-go client wiring (search, images, scrape, proxy)
pkg/ reusable building blocks (mcpserver, server, logger, validator)Flujo de solicitud para una llamada de herramienta:
MCP client → transport/mcp/web (handler) → usecase/web → adapter/web → retrieval-go → the web
↑ maps errors ↑ validates, fans out, limits resultsTanto la búsqueda como el raspado se distribuyen en paralelo a través de la lista de entrada y agregan resultados por elemento, cada uno con su propio estado (success, failed, timeout). Una llamada solo falla por completo cuando todos los elementos de ella fallan.
Motor de recuperación
Todo el trabajo de red se delega en retrieval-go, configurado en app/internal/adapter/web. Vale la pena saber:
Fuentes. La búsqueda web usa DuckDuckGo Lite; la búsqueda de imágenes usa Bing Images; la obtención de páginas procesa el HTML sin procesar mediante un extractor de legibilidad y convierte el artículo principal a Markdown (tablas incluidas). No se requieren claves de API de motores de búsqueda.
Suplantación de navegador. El adaptador habilita
WithBrowserRotation(), por lo que cada solicitud se envía desde uno de ~11 perfiles de navegador reales elegidos al azar. Cada perfil combina una huella TLS/JA3 genuina (mediante uTLS) con unUser-Agenty encabezados de sugerencia de cliente coincidentes: Chrome 133/131/120 (Windows/macOS/Linux), Edge 131, Firefox 120 (Windows/macOS), Safari 18.4 (macOS) y Safari de iOS 18.4. Esto hace que el tráfico parezca de navegadores ordinarios en lugar de un cliente HTTP de Go, que es lo que mantiene accesibles las fuentes gratuitas.Rotación de proxy. Cuando se configura
PROXY_HOST, el adaptador instala una fábrica de proxies que añade unsession-<id>único al nombre de usuario del proxy en cada solicitud. Con un proveedor de proxy residencial/rotatorio basado en sesiones, esto produce una IP de salida nueva por solicitud, distribuyendo la carga y evitando límites de velocidad. Sin proxy, las solicitudes salen directamente.Manejo de respuestas. Las respuestas se descomprimen de forma transparente (
gzip,br,zstd,deflate), y keep-alive está deshabilitado (WithDisableKeepAlive()) para que las conexiones agrupadas no fijen una sola huella/IP entre solicitudes.
Nada de esto necesita configuración para funcionar: los valores predeterminados anteriores se aplican automáticamente. Solo las credenciales del proxy son extras opcionales.
Desarrollo
Todo lo relacionado con Go vive en app/, así que usa el makefile desde la raíz del repositorio o pasa -C app a la cadena de herramientas:
make build # compile the binary
make test # run tests
go -C app build ./... # compile everything
go -C app test ./... # run tests
go -C app vet ./... # static checksConsulta CONTRIBUTING.md para las pautas de solicitudes de extracción.
Licencia
Publicado bajo la Licencia MIT.
Available Tools
3 toolsweb_scrapeWeb scrapeARead-onlyIdempotent
Downloads pages by their links and returns the main text as markdown. Use this to read a page in full after 'web_search' found it; snippets alone are often not enough. Links are fetched in parallel, up to 10 per call, so batch all links into one call.
Size: 'max_chars' is a per-page character limit (not tokens). A page cut short ends with '[truncated]' and has 'truncated' set; the rest cannot be fetched afterwards.
Results are reported per URL with a 'status' field (success, failed or timeout). A failed URL does not fail the others; the call errors only if every URL fails. Target sites may block or rate-limit the request, which shows up as a failed URL.
| Name | Required | Description | Default |
|---|---|---|---|
| urls | Yes | 1-10 absolute http(s) page URLs, fetched in parallel; one malformed URL rejects the whole call | |
| max_chars | No | character limit applied to each page separately, not to the whole response; default and maximum 20000, so it can only lower the limit; a cut page ends with '[truncated]' | |
| robots_txt | No | if true, pages disallowed by the site robots.txt are not fetched; default false | |
| timeout_ms | No | timeout for the whole call in milliseconds, shared by all URLs; default 5000, clamped to 1000-10000; unfinished pages get status 'timeout' | |
| remove_links | No | if true, markdown links are replaced by their text; default false |
Output Schema
| Name | Required | Description |
|---|---|---|
| results | Yes | |
| metadata | Yes |
TDQS
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare the safety profile (readOnly, idempotent, openWorld, non-destructive), yet the description still adds real substance: per-URL status values, partial-failure semantics ('a failed URL does not fail the others'), all-fail error condition, external blocking/rate-limiting, and the irreversibility of truncation. Minor gap: it doesn't mention the robots_txt opt-in or the call-wide timeout being shared across URLs, both of which affect real call behavior.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Front-loaded with the core action and the batching rule, then cleanly separated into size and result-reporting sections. Every sentence carries a constraint or a consequence an agent would otherwise get wrong; nothing is padding.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Covers the things that actually break an agent's plan: batching limit, per-page truncation, partial failure isolation, and timeouts. An output schema exists, yet the description still summarizes the per-URL status contract, which is a useful addition rather than redundancy.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 100%, so the baseline is 3, but the description adds interpretive value beyond the schema: 'max_chars' is a per-page limit, not tokens, and a cut page cannot be recovered afterwards. The remaining parameters (robots_txt, timeout_ms, remove_links) are left entirely to the schema, which is acceptable given full coverage.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
States a specific verb and resource ('Downloads pages by their links and returns the main text as markdown') and explicitly positions itself against the sibling 'web_search' by naming the gap in snippets. An agent can distinguish it from web_search and web_search_images 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.
Does the description explain when to use this tool, when not to, or what alternatives exist?
Gives an explicit sequencing rule ('use this to read a page in full after web_search found it; snippets alone are often not enough') and an operational rule ('batch all links into one call'). The alternative tool is named together with the condition that selects it.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
web_searchWeb searchARead-onlyIdempotent
Searches the web for one or more queries and returns snippets with links. Queries run in parallel. IMPORTANT: Always write search queries in English for best results and relevance. Use this to find sources; then call 'web_scrape' to read a full page, or 'web_search_images' when you need pictures (both can be called in the same turn).
Freshness: set the 'date' field to restrict results by recency - 'd' (past day), 'w' (past week), 'm' (past month), 'y' (past year). Use it to prefer the most recent pages when the user asks about news or anything time-sensitive; leave it empty for all time.
Query operators (put them inside the query string itself): 'site:example.com term' limits the search to one site; 'filetype:pdf term' restricts to a file type; double quotes "exact phrase" force an exact match; a leading '-' excludes a word (term -foo); 'intitle:term' requires the word in the page title. Operators can be combined, e.g. 'site:arxiv.org filetype:pdf transformers'.
Results are reported per query with a 'status' field (success, failed or timeout). A failed query does not fail the others; the call errors only if every query fails. The upstream search engines may rate-limit or block requests, which shows up as a failed or timed-out query - retry later or rephrase.
| Name | Required | Description | Default |
|---|---|---|---|
| date | No | freshness filter: 'd' past day, 'w' past week, 'm' past month, 'y' past year; empty means all time; use only these values | |
| queries | Yes | 1-10 search queries in English, run in parallel; each must be non-empty and at most 512 characters, otherwise the whole call is rejected | |
| timeout_ms | No | timeout for the whole call in milliseconds, shared by all queries; default 5000, clamped to 1000-10000; unfinished queries get status 'timeout' | |
| max_results | No | maximum snippets per query, not in total; default 5, values above 20 are clamped to 20 |
Output Schema
| Name | Required | Description |
|---|---|---|
| results | Yes | |
| metadata | Yes |
TDQS
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Goes well beyond the readOnly/openWorld/idempotent annotations: it explains per-query status reporting (success/failed/timeout), that a failed query does not fail the others and the call errors only if all queries fail, and that upstream rate-limiting surfaces as failed/timed-out queries requiring retry or rephrasing. That is exactly the operational context annotations cannot carry.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Front-loaded with purpose before the operational detail, and section labels (Freshness, Query operators) make it scannable. Minor redundancy: the date filter values are spelled out in both the description and the schema's date property description, costing a sentence that adds no new information.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Covers everything an agent needs for a multi-query search tool: batching behavior, partial-failure semantics, timeout/rate-limit handling, query formatting, and result shape at a high level. An output schema exists, so return values need not be enumerated here.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Even though schema coverage is 100%, the description adds substantial meaning: it mandates English queries, enumerates the supported query operators (site:, filetype:, quoted phrases, leading '-', intitle:) with a combined example, and explains freshness semantics. This is syntax guidance the schema does not express.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
States a specific verb and resource ('searches the web') plus scope ('one or more queries', parallel execution, snippets with links). It explicitly distinguishes itself from both siblings by naming web_scrape for full-page reads and web_search_images for pictures.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
Gives explicit when-to-use routing ('Use this to find sources; then call web_scrape...') and notes both can be called in the same turn. It also supplies conditional guidance for the date filter ('use it when the user asks about news or anything time-sensitive; leave it empty for all time').
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
web_search_imagesWeb image searchARead-onlyIdempotent
Searches for images by one or more queries, executed in parallel. IMPORTANT: Queries must be in English ONLY. Use this only when you need pictures; use 'web_search' for text answers and links, and 'web_scrape' to read a page.
Freshness: set the 'date' field to restrict results by recency - 'd' (past day), 'w' (past week), 'm' (past month), 'y' (past year); leave it empty for all time.
Query operators can be embedded in the query string: 'site:example.com term' limits to one site, double quotes "exact phrase" force an exact match, and a leading '-' excludes a word.
Results are reported per query with a 'status' field (success, no relevant, failed or timeout). A failed query does not fail the others; the call errors only if every query fails. The upstream image search may rate-limit or block requests, which shows up as a failed or timed-out query.
| Name | Required | Description | Default |
|---|---|---|---|
| date | No | freshness filter: 'd' past day, 'w' past week, 'm' past month, 'y' past year; empty means all time; other values are ignored | |
| queries | Yes | 1-10 image search queries in English, run in parallel; each must be non-empty and at most 512 characters, otherwise the whole call is rejected | |
| max_images | No | maximum images per query, not in total; default 5, values above 10 are clamped to 10 | |
| timeout_ms | No | timeout for the whole call in milliseconds, shared by all queries; default 5000, clamped to 1000-10000; unfinished queries get status 'timeout' |
Output Schema
| Name | Required | Description |
|---|---|---|
| results | Yes | |
| metadata | Yes |
TDQS
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare readOnly/openWorld/idempotent/non-destructive, yet the description still adds real value: per-query status semantics, partial-failure behavior ("a failed query does not fail the others"), and the upstream rate-limit/block risk. It does not explain result ordering or what fields come back, but the output schema covers that.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Front-loaded with the core behavior and routing decision, then broken into scannable paragraphs for freshness, operators, and failure semantics. Slightly dense, but no sentence is filler.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a 4-param parallel search tool with annotations and an output schema, this covers everything an agent needs: language constraint, batching, freshness, operator syntax, and failure modes. Nothing material is missing.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 100%, so baseline is 3, and the description goes beyond it by documenting query operators (site:, "exact phrase", leading '-') that are absent from the schema entirely. The freshness codes and clamping are repeated from the schema rather than adding to it.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
States a specific verb and resource ("Searches for images by one or more queries") plus the parallel-execution trait. It explicitly names both siblings and what each is for, so an agent can pick the right tool without opening a schema.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
Gives an explicit use condition ("use this only when you need pictures") and names the two alternatives with their respective purposes. The English-only constraint is flagged as IMPORTANT and front-loaded.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
3 tool updates
v0.1.10- Changed
web_scrape6 fields changed- changed
Input schema / properties / max_chars / descriptionPrevious value: -"truncate each page separately to N characters, default and maximum 20000; only lowers the limit, higher values are ignored"New value: +"character limit applied to each page separately, not to the whole response; default and maximum 20000, so it can only lower the limit; a cut page ends with '[truncated]'" - changed
Input schema / properties / remove_links / descriptionPrevious value: -"strip markdown links from the text"New value: +"if true, markdown links are replaced by their text; default false" - changed
Input schema / properties / robots_txt / descriptionPrevious value: -"respect the page robots.txt"New value: +"if true, pages disallowed by the site robots.txt are not fetched; default false" - changed
Input schema / properties / timeout_ms / descriptionPrevious value: -"timeout for the whole call in milliseconds"New value: +"timeout for the whole call in milliseconds, shared by all URLs; default 5000, clamped to 1000-10000; unfinished pages get status 'timeout'" - changed
Input schema / properties / urls / descriptionPrevious value: -"links to pages, downloaded in parallel"New value: +"1-10 absolute http(s) page URLs, fetched in parallel; one malformed URL rejects the whole call" - added
Input schema / properties / urls / maxItemsAdded value: +10
- Changed
web_search5 fields changed- changed
Input schema / properties / date / descriptionPrevious value: -"filter results by freshness: 'd' past day, 'w' past week, 'm' past month, 'y' past year; empty means all time"New value: +"freshness filter: 'd' past day, 'w' past week, 'm' past month, 'y' past year; empty means all time; use only these values" - changed
Input schema / properties / max_results / descriptionPrevious value: -"maximum number of snippets per query"New value: +"maximum snippets per query, not in total; default 5, values above 20 are clamped to 20" - changed
Input schema / properties / queries / descriptionPrevious value: -"search queries in English, executed in parallel"New value: +"1-10 search queries in English, run in parallel; each must be non-empty and at most 512 characters, otherwise the whole call is rejected" - added
Input schema / properties / queries / maxItemsAdded value: +10 - changed
Input schema / properties / timeout_ms / descriptionPrevious value: -"timeout for the whole call in milliseconds"New value: +"timeout for the whole call in milliseconds, shared by all queries; default 5000, clamped to 1000-10000; unfinished queries get status 'timeout'"
- Changed
web_search_images5 fields changed- changed
Input schema / properties / date / descriptionPrevious value: -"filter results by freshness: 'd' past day, 'w' past week, 'm' past month, 'y' past year; empty means all time"New value: +"freshness filter: 'd' past day, 'w' past week, 'm' past month, 'y' past year; empty means all time; other values are ignored" - changed
Input schema / properties / max_images / descriptionPrevious value: -"maximum number of images per query"New value: +"maximum images per query, not in total; default 5, values above 10 are clamped to 10" - changed
Input schema / properties / queries / descriptionPrevious value: -"queries in English for image search"New value: +"1-10 image search queries in English, run in parallel; each must be non-empty and at most 512 characters, otherwise the whole call is rejected" - added
Input schema / properties / queries / maxItemsAdded value: +10 - changed
Input schema / properties / timeout_ms / descriptionPrevious value: -"timeout for the whole call in milliseconds"New value: +"timeout for the whole call in milliseconds, shared by all queries; default 5000, clamped to 1000-10000; unfinished queries get status 'timeout'"
3 tool updates
v1.0.0- First observed
web_scrape - First observed
web_search - First observed
web_search_images
TDQS
Scored across 3 tools
Each tool targets a distinct modality+action: text search, image search, and full-page fetching. The descriptions explicitly cross-reference each other ('use web_search for text, web_scrape to read a page'), leaving no realistic chance of misselection.
All three names follow the same snake_case verb_noun pattern sharing the 'web_' prefix (web_search, web_search_images, web_scrape). The shared prefix makes the family immediately recognizable while the distinguishing suffix stays descriptive.
Three tightly scoped tools for a retrieval-only server; none is redundant and no filler exists. Search-text, search-images, and scrape cover the full surface without bloat.
The retrieve-then-read workflow is fully supported, with batching, truncation signaling, and per-URL status handling. Minor gaps remain (no structured/raw fetch beyond markdown extraction, no crawl or sitemap traversal), but agents can work around these via repeated scrape calls.
Maintenance
Related MCP Connectors
Free web search for AI agents. No API key required. Hosted MCP in active development.
Docs: https://docs.keenable.ai/mcp-server Keenable is a free, remote MCP server that gives agents access to the web index. Search the web with ranked results and date/site filters, then fetch any indexed page as clean markdown. Works out of the box with no account or API key.
One MCP server for all web data. Scrape, search, crawl, and AI research in one install.
Related MCP Servers
- AlicenseBqualityDmaintenanceA local web scraping MCP server with RAG capabilities that provides intelligent web search, content extraction, and screenshot tools without requiring API keys.46 npmMIT
- AlicenseAqualityDmaintenanceAn MCP server that enables web searching, URL content extraction, and summarization without requiring API keys. It also provides advanced mathematical evaluation and multi-language Wikipedia summary retrieval tools.5400 npm7MIT
- AlicenseAqualityAmaintenanceA local-first, no-API-key MCP server that enables LLMs to search the web, fetch pages, and read documents using multiple engines and smart fallbacks.11167MIT
- AlicenseAqualityAmaintenanceA self-hosted MCP server providing web search and URL fetching tools, running locally without external API keys or accounts.2380 npmMIT