Skip to content
Factflow

Search Vector Only

POST
/api/v1/chat/search/vector

Execute vector-only semantic search.

Uses embedding similarity without keyword matching. Useful when exact terms don’t matter, only semantic meaning.

Args: request: Vector search request service: Hybrid search service (injected)

Returns: Search results ranked by vector similarity

Raises: HTTPException: 503 if service not configured

Request Body required

Section titled “Request Body required ”
VectorSearchRequest

Request model for vector-only search.

object
query
required
Query

Search query text

string
>= 1 characters <= 10000 characters
model_name
required
Model Name

Embedding model name

string
execution_id
Any of:
string format: uuid
limit
Limit

Maximum results to return

integer
default: 10 >= 1 <= 100
similarity_threshold
Similarity Threshold

Minimum similarity score

number
0 <= 1

Responses

Section titled “ Responses ”

200

Section titled “200 ”

Successful Response

RRFSearchResponse

Response model for RRF hybrid search.

object
results
required
Results
Array<object>
RRFResultModel

Single result from RRF hybrid search.

object
segment_id
required
Segment Id
string format: uuid
rrf_score
required
Rrf Score
number
content
required
Content
string
storage_key
required
Storage Key
string
language
required
Language
string
vector_similarity
required
Any of:
number
keyword_rank
required
Any of:
number
vector_rrf_contribution
required
Vector Rrf Contribution
number
keyword_rrf_contribution
required
Keyword Rrf Contribution
number
total_candidates
required
Total Candidates
integer
language
required
Language
string
search_time_ms
required
Search Time Ms
number
config
required
Config
object
key
additional properties
any

422

Section titled “422 ”

Validation Error

HTTPValidationError
object
detail
Detail
Array<object>
ValidationError
object
loc
required
Location
Array
msg
required
Message
string
type
required
Error Type
string
input
Input
ctx
Context
object