Documentation Index Fetch the complete documentation index at: https://docs.bedrock.cv/llms.txt
Use this file to discover all available pages before exploring further.
Prerequisites
A Bedrock account with an API key
A construction drawing PDF
Set your API key:
export BEDROCK_API_KEY = sk_xxx
1. Create a Project curl -X POST https://api.bedrock.cv/projects \ -H "Authorization: Bearer $BEDROCK_API_KEY " \ -H "Content-Type: application/json" \ -d '{"name": "My First Project"}'
{ "id" : "prj_01JABCD123" , "name" : "My First Project" , "created_at" : "2024-01-15T10:00:00Z" }
2. Upload a Drawing Upload a PDF file, then create a drawing from it.
# Create a file upload curl -X POST https://api.bedrock.cv/files \ -H "Authorization: Bearer $BEDROCK_API_KEY " \ -H "Content-Type: application/json" \ -d '{ "project_id": "prj_01JABCD123", "filename": "arch-set-rev-b.pdf", "content_type": "application/pdf" }'
{ "id" : "fil_01JABCD111" , "upload_url" : "https://storage.bedrock.cv/upload/..." , "upload_headers" : { "Content-Type" : "application/pdf" } }
# Upload the PDF to the signed URL curl -X PUT "https://storage.bedrock.cv/upload/..." \ -H "Content-Type: application/pdf" \ --data-binary @arch-set-rev-b.pdf # Create a drawing from the uploaded file curl -X POST https://api.bedrock.cv/drawings \ -H "Authorization: Bearer $BEDROCK_API_KEY " \ -H "Content-Type: application/json" \ -d '{ "project_id": "prj_01JABCD123", "file_id": "fil_01JABCD111", "name": "Architectural Set Rev B" }'
{ "id" : "drw_01JABCD222" , "name" : "Architectural Set Rev B" , "job_id" : "job_01JABCD333" , "created_at" : "2024-01-15T10:01:00Z" }
3. Wait for Processing Bedrock’s vision pipeline automatically splits the PDF into sheets, detects blocks, and parses features. Poll the job until it completes:
curl https://api.bedrock.cv/jobs/job_01JABCD333 \ -H "Authorization: Bearer $BEDROCK_API_KEY "
{ "job_id" : "job_01JABCD333" , "type" : "drawing.preprocess" , "status" : "Completed" , "created_at" : "2024-01-15T10:01:00Z" , "completed_at" : "2024-01-15T10:02:30Z" }
4. Query the Drawing Index Your drawing is now parsed into the Drawing Index. Here are two ways to work with it:
List sheets in your drawing:curl "https://api.bedrock.cv/sheets?drawing_id=drw_01JABCD222" \ -H "Authorization: Bearer $BEDROCK_API_KEY "
{ "data" : [ { "id" : "sht_01A" , "sheet_number" : "A-101" , "title" : "First Floor Plan" , "discipline" : "A" }, { "id" : "sht_01B" , "sheet_number" : "A-102" , "title" : "Second Floor Plan" , "discipline" : "A" }, { "id" : "sht_01C" , "sheet_number" : "A-501" , "title" : "Details" , "discipline" : "A" } ] }
Query features — count duplex receptacles on a sheet:curl "https://api.bedrock.cv/features?sheet_number=E-201&type=duplex_receptacle&project_id=prj_01JABCD123" \ -H "Authorization: Bearer $BEDROCK_API_KEY "
{ "data" : [ { "id" : "ftr_01A" , "type" : "duplex_receptacle" , "label" : null , "parent_feature_id" : "ftr_room201" }, { "id" : "ftr_01B" , "type" : "duplex_receptacle" , "label" : null , "parent_feature_id" : "ftr_room201" }, { "id" : "ftr_01C" , "type" : "duplex_receptacle" , "label" : null , "parent_feature_id" : "ftr_room202" } ], "_meta" : { "total" : 17 } }
Get a feature with its relations:curl "https://api.bedrock.cv/features/ftr_door104?expand=relations,block" \ -H "Authorization: Bearer $BEDROCK_API_KEY "
See the full CMS API reference for all endpoints. Connect the Drawing Index to your AI agent and query with natural language. Add to your MCP configuration: { "mcpServers" : { "bedrock" : { "type" : "http" , "url" : "https://mcp.bedrock.cv/mcp" , "headers" : { "Authorization" : "Bearer sk_xxx" } } } }
Then ask your agent:
“How many duplex receptacles are on sheet E-201?”
The agent calls query_feature and returns:
17 duplex receptacles on E-201: 4 in Room 201, 3 in Room 202, 2 in Room 203…
“What are the specs for door 104?”
The agent calls query_feature to find the door, follows the scheduled_in relation to the Door Schedule, and returns:
Door 104: 3’-0” x 7’-0”, hollow metal, 45-min fire rating, hardware set HW-3.
See the full MCP setup guide for Claude, ChatGPT, and Microsoft Copilot configuration. Next Steps
Core Concepts Understand the Drawing Index data model — hierarchy, features, relations, and grids.
MCP Workflows See how AI agents chain tools to answer construction questions.
