Skip to main content
POST
/
projects
/
{project_id}
/
ask
Ask a question about a project
curl --request POST \
  --url https://api.bedrock.cv/projects/{project_id}/ask \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "question": "<string>"
}
'
{
  "answer": "<string>",
  "sources": [
    {
      "source_type": "block",
      "source_id": "<string>",
      "label": "<string>",
      "score": 123
    }
  ]
}
Ask answers “what does this project say about X?” by searching across all project content (blocks, features, files), then generating a concise answer grounded in the retrieved context. Ask is synchronous and returns results immediately (not an async job).

How It Works

  1. Your question is used to run a hybrid search (semantic + keyword) across the project
  2. The top results are assembled into a context window
  3. An LLM generates a 2-4 sentence answer citing the source material
  4. Up to 5 source references are returned with relevance scores

Example Questions

QuestionWhat it finds
”What fire rating is required for the stairwell doors?”Spec blocks, general notes referencing fire ratings
”How many parking spaces are shown?”Site plan blocks with parking counts
”What concrete mix is specified for the foundation?”Specification sheets, structural notes

Sources

Each source in the response includes:
FieldDescription
source_typeType of content matched (block, feature, or file)
source_idID of the matched entity
labelHuman-readable label (e.g. sheet number, file name)
scoreRelevance score (higher is better)

Authorizations

Authorization
string
header
required

API key prefixed with sk_. Example: Authorization: Bearer sk_xxx

Headers

X-API-Version
string
default:2026-01-01

API version

Path Parameters

project_id
string
required

Project ID

Body

application/json
question
string
required

Response

Generated answer with sources.

answer
string

Generated answer to the question

sources
object[]