RAG APIs
RAG Pipeline — Chunk, Index, Embed & Search
A complete set of APIs to build a Retrieval-Augmented Generation (RAG) pipeline. Use them to chunk and index a PDF document, perform keyword, vector, and semantic search over indexed content, and generate grounded answers from enterprise documents using Azure OpenAI GPT.
Postman Collection Variables — Set Before Use
azureAiSearch_ResourceName
azureAISearch_IndexName
azureAISearch_Endpoint
azureAISearch_Key
openAI_ResourceName
openAI_Key
openAI_EmbeddingModelName
openAI_GPTModelName
openAI_Endpoint
The collection uses {{variable}} placeholders throughout. Open the collection in Postman, go to Variables, and fill in your Azure AI Search and Azure OpenAI credentials before running any request. APIs 01 and 08 call the Lowcademy-hosted endpoint — the rest call Azure directly.
| # | API Name | Method | Auth | View Docs |
|---|---|---|---|---|
| 01 |
Chunk & Perform Embedding
labs.lowcademy.com/apis/chunk-and-perform-embedding.php
Upload PDF → chunk → embed → Azure AI Search payload
|
POST | Lowcademy Key | View → |
| 02 |
Index Document
{resource}.search.windows.net/indexes/{index}/docs/index
POST Step 1 response directly to Azure AI Search
|
POST | Azure Search Key | View → |
| 03 |
Text Search (Keyword)
{resource}.search.windows.net/indexes/{index}/docs/search
BM25 full-text keyword retrieval
|
POST | Azure Search Key | View → |
| 04 |
Semantic Search
{resource}.search.windows.net/indexes/{index}/docs/search
Meaning-based L2 re-ranking with semantic captions
|
POST | Azure Search Key | View → |
| 05 |
Get Question Embedding
{openAI_Endpoint}openai/deployments/{model}/embeddings
Convert question to 1536-dim vector for vector search
|
POST | Azure OpenAI Key | View → |
| 06 |
Vector Search
{resource}.search.windows.net/indexes/{index}/docs/search
ANN similarity search using the question embedding vector
|
POST | Azure Search Key | View → |
| 07 |
Ask Question to GPT
{openAI_Endpoint}openai/deployments/{model}/chat/completions
Send retrieved chunks as context → GPT generates grounded answer
|
POST | Azure OpenAI Key | View → |
| 08 |
Delete Indexed Documents
labs.lowcademy.com/apis/delete-indexed-document.php
Remove all Azure AI Search chunks by document_id
|
POST | Lowcademy + Azure | View → |