Search documentation

Find canonical links and titles in the published GoVisit documentation index. Use the bounded matches to choose the page you want to read next.

What it returns

Search is limited to documentation links rendered on the current page. Follow canonical links for the full document.

For the full current contract and linked context, see the WebMCP manifest and the WebMCP overview.

WebMCP

Input schema

{
  "type": "object",
  "properties": {
    "query": {
      "type": "string",
      "description": "Optional documentation title or path text to match."
    },
    "limit": {
      "type": "integer",
      "minimum": 1,
      "maximum": 20
    }
  },
  "additionalProperties": false
}

Output schema

{
  "type": "object",
  "properties": {
    "data": {
      "type": "object",
      "properties": {
        "items": {
          "type": "array"
        },
        "count": {
          "type": "integer",
          "minimum": 0
        }
      },
      "required": [
        "items",
        "count"
      ]
    }
  },
  "required": [
    "data"
  ]
}

const modelContext = document.modelContext;
if (!modelContext) throw new Error("This browser does not expose WebMCP on this page.");
const tool = (await modelContext.getTools()).find((entry) => entry.name === 'search_documentation');
if (!tool) throw new Error('search_documentation is not registered for the current page.');
const result = await modelContext.executeTool(tool, {"query": "WebMCP", "limit": 5});
console.log(result);

Registration is page-scoped. Check getTools() before calling.

HTTP API

Request body schema

{
  "type": "object",
  "properties": {
    "pagePath": {
      "type": "string",
      "minLength": 1,
      "maxLength": 1000,
      "description": "Canonical GoVisit page path supplying this tool's published context; no origin, query or fragment."
    },
    "query": {
      "type": "string",
      "description": "Optional documentation title or path text to match."
    },
    "limit": {
      "type": "integer",
      "minimum": 1,
      "maximum": 20
    }
  },
  "additionalProperties": false,
  "required": [
    "pagePath"
  ]
}

Successful response schema

{
  "type": "object",
  "properties": {
    "data": {
      "type": "object",
      "properties": {
        "items": {
          "type": "array"
        },
        "count": {
          "type": "integer",
          "minimum": 0
        }
      },
      "required": [
        "items",
        "count"
      ]
    }
  },
  "required": [
    "data"
  ]
}

Error response schema

{
  "type": "object",
  "properties": {
    "schemaVersion": {
      "const": "1"
    },
    "completeness": {
      "enum": [
        "invalid-input",
        "not-found",
        "unavailable"
      ]
    },
    "error": {
      "type": "string"
    }
  },
  "required": [
    "schemaVersion",
    "completeness",
    "error"
  ],
  "additionalProperties": false
}

Send a JSON POST to https://govisit.ai/api/tools/v1/search_documentation. See the HTTP API guide for shared conventions.

Use the pathname from an actual canonical GoVisit link, including its language prefix. The sample page path illustrates the shape; replace it with your selected published page. HTTP inputs add pagePath to the browser arguments. No cookies, browser session or inferred location are required. Queries, fragments, origins and cache-refresh flags are rejected. Language comes from the page path.

curl

curl --silent --show-error https://govisit.ai/api/tools/v1/search_documentation \
  -H 'Content-Type: application/json' \
  --data '{"pagePath": "/en/docs", "query": "Developers", "limit": 5}'

JavaScript

const response = await fetch("https://govisit.ai/api/tools/v1/search_documentation", {
  method: "POST",
  headers: { "Content-Type": "application/json" },
  body: JSON.stringify({
  "pagePath": "/en/docs",
  "query": "Developers",
  "limit": 5
})
});
const result = await response.json();
if (!response.ok) throw new Error(`${response.status}: ${result.error || "unavailable"}`);
console.log(result);

Python

import json
from urllib.error import HTTPError
from urllib.request import Request, urlopen

request = Request(
    "https://govisit.ai/api/tools/v1/search_documentation",
    data=json.dumps({
  "pagePath": "/en/docs",
  "query": "Developers",
  "limit": 5
}).encode(),
    headers={"Content-Type": "application/json"},
    method="POST",
)
try:
    with urlopen(request) as response:
        result = json.load(response)
except HTTPError as error:
    result = json.load(error)
    raise RuntimeError(f"{error.code}: {result.get('error', 'unavailable')}") from error
print(result)

Successful results match the browser tool's published page projection, including its freshness and scope limits. The adapter uses GoVisit's existing page caches; it does not provide a fresh live availability check. An absent page returns 404; an unsupported section returns 400; unavailable or malformed page context returns 503. See Reliability for errors and shared request limits.

MCP

Tool arguments schema

{
  "type": "object",
  "properties": {
    "pagePath": {
      "type": "string",
      "minLength": 1,
      "maxLength": 1000,
      "description": "Canonical GoVisit page path supplying this tool's published context; no origin, query or fragment."
    },
    "query": {
      "type": "string",
      "description": "Optional documentation title or path text to match."
    },
    "limit": {
      "type": "integer",
      "minimum": 1,
      "maximum": 20
    }
  },
  "additionalProperties": false,
  "required": [
    "pagePath"
  ]
}

Structured result schema

{
  "type": "object",
  "properties": {
    "data": {
      "type": "object",
      "properties": {
        "items": {
          "type": "array"
        },
        "count": {
          "type": "integer",
          "minimum": 0
        }
      },
      "required": [
        "items",
        "count"
      ]
    }
  },
  "required": [
    "data"
  ]
}

Connect a Streamable HTTP MCP client to https://govisit.ai/mcp and call search_documentation. Use the HTTP API arguments shown above, including the canonical pagePath. Replace example paths with the published page you want to read.

const result = await client.callTool({
  name: "search_documentation",
  arguments: {"pagePath": "/en/docs", "query": "Developers", "limit": 5},
});
if (result.isError) throw new Error(result.content[0]?.text ?? "GoVisit tool call failed");
console.log(result.structuredContent);

Agent Skills

Tool input schema

{
  "type": "object",
  "properties": {
    "pagePath": {
      "type": "string",
      "minLength": 1,
      "maxLength": 1000,
      "description": "Canonical GoVisit page path supplying this tool's published context; no origin, query or fragment."
    },
    "query": {
      "type": "string",
      "description": "Optional documentation title or path text to match."
    },
    "limit": {
      "type": "integer",
      "minimum": 1,
      "maximum": 20
    }
  },
  "additionalProperties": false,
  "required": [
    "pagePath"
  ]
}

Tool output schema

{
  "type": "object",
  "properties": {
    "data": {
      "type": "object",
      "properties": {
        "items": {
          "type": "array"
        },
        "count": {
          "type": "integer",
          "minimum": 0
        }
      },
      "required": [
        "items",
        "count"
      ]
    }
  },
  "required": [
    "data"
  ]
}

After installing GoVisit skills, ask your assistant to use govisit-search-documentation. The skill calls the GoVisit HTTP API.

Open the GoVisit documentation index in your preferred language. Name the topic you want to find. Search supports the localized index (/en/docs or /cs/docs); use Read documentation page to read an individual article.

Use GoVisit to search the documentation links on https://govisit.ai/en/docs for [topic]. Show matching articles and their source links.

Skill package and contract

Skill source · Complete package · HTTP contract

The complete package includes its MIT license and versioned contract. GoVisit data remains subject to the GoVisit Data Use License.