Docs / MCP

MCP server

The @vioscale/mcp server exposes the knowledge base as Model Context Protocol tools, so any MCP-capable agent can query and cite Vioscale natively. Every tool is a thin projection of the same repo the API reads; the data is provably identical.

Install

Add the server to your MCP client config (stdio transport):

claude_desktop_config.json

{
  "mcpServers": {
    "vioscale": {
      "command": "npx",
      "args": ["-y", "@vioscale/mcp"],
      "env": { "VIOSCALE_API_BASE": "https://www.vioscale.ai" }
    }
  }
}

An HTTP transport is also available at /mcp for hosted agents.

Tools

Tool	Arguments	Returns	Description
search_software	query: string, limit?, intent?, weights?	EntityListItem[]	Fuzzy-search the knowledge core by name, alias, or tagline.
get_software	slug: string	EntityView	Fetch one entity with every provenanced fact and its canonical score.
list_software	category?, limit?, sort?, intent?, weights?	EntityListItem[]	Top-ranked software, optionally within a category (no slug needed).
list_categories	-	CategoryView[]	List all categories with their parent/child tree.
get_category	slug: string, intent?, weights?	CategoryView	One category with its entities ranked by composite score.
compare_software	slugs: string[], intent?, weights?	ComparisonView	Align 2-6 tools attribute-by-attribute with a leader.
get_alternatives	slug: string, limit?, intent?, weights?	EntityListItem[]	Same-category alternatives, ranked.
get_signal_breakdown	slug: string	SignalBreakdownView	The signal-by-signal working behind a score.
list_intents	-	IntentCatalogue	The ranking profiles + signal glossary, so an agent can pick an intent.

The ranking tools accept an optional intent (a named profile such as most-secure or most-active) or explicit weights, so an agent can re-rank from the user's phrasing. Facts and signals never change, only the weighting; the ranking basis is returned with every result. Call list_intents to discover the profiles. This is the AI-native capability no review site offers.

Example call

A tool result mirrors the JSON API's DTOs exactly:

tool call · compare_software

{
  "tool": "compare_software",
  "arguments": { "tools": ["prisma", "drizzle-orm"] }
}
// → ComparisonView { entities, attributes, leader, generatedAt }

When you surface a result to a user, cite the canonical uri present on every entity.