veryfront__search_files
Requires project.knowledge.read.
Access
No external-service connection is required. Use a Veryfront credential with access to the requested resources.
Call type: Read-only.
Arguments
Example request
Send this JSON-RPC request through an initialized MCP client. Replace sample identifiers and values before calling the tool. See MCP client setup.
{
"jsonrpc": "2.0",
"id": 1,
"method": "tools/call",
"params": {
"name": "veryfront__search_files",
"arguments": {
"query_vector": [],
"dimension": "768"
}
}
}Result
The published output shape is summarized below. Inspect both the MCP response content and isError when handling a call.
Full schemas
{
"type": "object",
"properties": {
"project_reference": {
"type": "string",
"description": "Project ID or slug whose indexed file chunks should be searched."
},
"model": {
"type": "object",
"properties": {
"name": {
"type": "string",
"minLength": 1,
"description": "Human-readable name for the target record."
},
"provider": {
"type": "string",
"minLength": 1,
"description": "Provide the provider."
}
},
"required": [
"name",
"provider"
],
"additionalProperties": false,
"description": "AI model id or alias. Call list_models for the ids this deployment serves; never invent a version suffix."
},
"query_vector": {
"type": "array",
"items": {
"type": "number",
"description": "One query vector item supplied to the tool."
},
"maxItems": 4096,
"description": "Embedding vector for the search query, generated with a compatible embedding model."
},
"dimension": {
"type": "string",
"enum": [
"768",
"1024",
"1536",
"3072",
"4096"
],
"description": "Embedding dimension that matches the supplied query vector."
},
"mode": {
"type": "string",
"enum": [
"SEMANTIC",
"KEYWORD",
"HYBRID"
],
"description": "Search mode. Only SEMANTIC (the default) is supported; KEYWORD and HYBRID are refused."
},
"environment": {
"type": "string",
"enum": [
"PREVIEW",
"PRODUCTION",
"RELEASE"
],
"description": "Index to search: the preview branch, the production deployment or a release."
},
"branch_id": {
"anyOf": [
{
"type": "string",
"format": "uuid",
"description": "Branch identifier. Use null for the main branch when the tool supports it."
},
{
"type": "null",
"description": "Branch identifier. Use null for the main branch when the tool supports it."
}
],
"description": "Preview branch ID to search. Use null for the main branch."
},
"release_id": {
"type": "string",
"format": "uuid",
"description": "Release ID to search instead of the current branch or mainline index."
},
"release_version": {
"type": "string",
"description": "Release version to search when you do not have the release ID."
},
"limit": {
"type": "number",
"minimum": 1,
"maximum": 100,
"description": "Maximum number of ranked chunk matches to return, as an integer from 1 to 100."
},
"cursor": {
"type": "string",
"description": "Opaque pagination cursor from a previous response."
},
"options": {
"type": "object",
"properties": {
"similarity_threshold": {
"type": "number",
"minimum": 0,
"maximum": 1,
"description": "Minimum similarity score required for a chunk to be returned."
},
"include_embeddings": {
"type": "boolean",
"description": "Include embedding vectors in results."
},
"folders": {
"type": "array",
"items": {
"type": "string",
"description": "One folder item supplied to the tool."
},
"description": "Reserved folder filter. A non-empty list is refused until folder filtering is supported."
},
"file_types": {
"type": "array",
"items": {
"type": "string",
"description": "One file type item supplied to the tool."
},
"description": "Reserved file-type filter. A non-empty list is refused until file-type filtering is supported."
}
},
"additionalProperties": false,
"description": "Search options. Tool results carry no embedding vectors, so include_embeddings has no effect here. Non-empty folders or file_types are refused: the search does not filter by them yet."
}
},
"required": [
"query_vector",
"dimension"
],
"additionalProperties": false,
"description": "Arguments for semantic file search over indexed project chunks.",
"$schema": "http://json-schema.org/draft-07/schema#"
}{
"type": "object",
"properties": {
"data": {
"type": "array",
"items": {
"type": "object",
"properties": {
"chunk": {
"type": "object",
"properties": {
"id": {
"type": "string",
"description": "Indexed chunk identifier."
},
"file_path": {
"type": "string",
"description": "Project-relative file path that produced this chunk."
},
"file_id": {
"type": "string",
"description": "File identifier when the source file is tracked as a project file."
},
"file_version_id": {
"type": "string",
"description": "File version identifier used to generate this chunk."
},
"index": {
"type": "number",
"description": "Zero-based chunk position within the file version."
},
"content": {
"type": "string",
"description": "Chunk text content used for retrieval and inspection."
},
"start_offset": {
"type": "number",
"description": "Inclusive character offset where the chunk begins in the source file."
},
"end_offset": {
"type": "number",
"description": "Exclusive character offset where the chunk ends in the source file."
},
"token_count": {
"type": "number",
"description": "Approximate token count for the chunk content."
},
"embedding": {
"anyOf": [
{
"type": "object",
"properties": {
"id": {
"type": "string",
"description": "Embedding record identifier for this indexed chunk."
},
"model": {
"type": "string",
"description": "Embedding model name used to create the vector."
},
"status": {
"type": "string",
"enum": [
"ready",
"stale",
"missing"
],
"description": "Embedding availability status for the chunk."
}
},
"required": [
"id",
"model",
"status"
],
"additionalProperties": false,
"description": "Embedding metadata attached to an indexed file chunk."
},
{
"type": "null",
"description": "The embedding returned by the tool."
}
],
"description": "Embedding metadata for the chunk, or null when no embedding is stored."
}
},
"required": [
"id",
"file_path",
"file_version_id",
"index",
"content",
"start_offset",
"end_offset",
"token_count",
"embedding"
],
"additionalProperties": false,
"description": "Matched file chunk with the most useful retrieval fields for ranking and display."
},
"score": {
"type": "number",
"description": "Similarity score between the query vector and the returned chunk."
}
},
"required": [
"chunk",
"score"
],
"additionalProperties": false,
"description": "One ranked chunk match returned by semantic file search."
},
"description": "Ranked semantic matches for the supplied query vector."
},
"page_info": {
"type": "object",
"properties": {
"self": {
"type": [
"string",
"null"
],
"description": "Cursor representing this result page."
},
"first": {
"type": "object",
"description": "Always null because semantic search pagination is forward-only."
},
"next": {
"type": [
"string",
"null"
],
"description": "Cursor for the next page of ranked matches, or null when exhausted."
},
"prev": {
"type": [
"string",
"null"
],
"description": "Cursor for the previous page when supported by the backend, otherwise null."
}
},
"required": [
"self",
"first",
"next",
"prev"
],
"additionalProperties": false,
"description": "Pagination cursors for traversing semantic search results."
},
"total_matches": {
"type": "number",
"description": "Total number of chunk matches that met the similarity threshold."
},
"search_time_ms": {
"type": "number",
"description": "Server-side search execution time in milliseconds."
}
},
"required": [
"data",
"page_info",
"total_matches",
"search_time_ms"
],
"additionalProperties": false,
"description": "Semantic file search results ranked by vector similarity.",
"$schema": "http://json-schema.org/draft-07/schema#"
}