top of page

Semantic search using AIWhispr with QDrant vector database

AIWhispr is a tool to enable AI powered semantic search on documents

  • It is easy to install.

  • Simple to configure.

  • Can handle multiple file formats (txt,csv, pdf, docx,pptx, docx) stored on AWS S3, Azure Blob Containers, local directory path.

  • Delivers fast semantic response to search queries.

  • Enables you to integrate with LLM's and vector databases of your choice; you can write you own custom modules or leverage the inbuilt modules in AIWhispr.

Adding support for a vector database is 2 simple steps, including

  1. Add a module for the vector database. To support Qdrant we have released a module

  2. Mention this module name in the [vectordb] section of the configuration file

Prerequisites for a Linux install with Qdrant as vector database

Environment variables

AIWHISPR_HOME_DIR environment variable should be the full path to aiwhispr directory.

AIWHISPR_LOG_LEVEL environment variable can be set to DEBUG / INFO / WARNING / ERROR


Download Qdrant and run the service

docker pull qdrant/qdrant 

Run the service

docker run -p 6333:6333 \     
       -v $(pwd)/qdrant_storage:/qdrant/storage \     

Qdrant should be accessible at localhost:6333

Python packages

Install python package for AIWhispr


Your first setup

1. Configuration file

A configuration file is maintained under $AIWHISPR_HOME/config/content-site/sites-available

We will use the example under examples/http to create a config file to index over 2000+ files which contain BBC news content.

To create the config file run the following commands.

You can enter "N" and choose to go with the default values

cd $AIWHISPR_HOME/examples/http;

It will display a config file that has been created.

#### CONFIG FILE ####
api-address= localhost
api-port= 6333

Check that config file has been created.

ls $AIWHISPR_HOME/config/content-site/sites-available/example_bbc.filepath.qdrant.cfg

2. Start Indexing

Confirm that the environment variables AIWHISPR_HOME and AIWHISPR_LOG_LEVEL are set and exported. Index the file content for semantic search. This will take some time as it has to process over 2000 files. The logging set to DEBUG will generate a verbose output.

$AIWHISPR_HOME/shell/ \ 
-C $AIWHISPR_HOME/config/content-site/sites-available/example_bbc.filepath.qdrant.cfg

This dataset has been sourced from

3. Start the AIWhispr search service

3 services will be started under this shell script.

  • the AIWhispr searchService(port:5002) which interfaces with the vectordb

  • a flask python script(port:9001) that takes in user query , sends the query to AIWhispr searchService and formats the results for HTML display

  • a python http.server(port 9000)

The log files for these 3 processes is created in /tmp/

cd $AIWHISPR_HOME/examples/http;

4. Ready for search

Try the search on

Try queries like

"What are the top TV moments in Olympics"

"Which is the best laptop to buy"

"How is inflation impacting the economy"

21 views0 comments

Recent Posts

See All

We continue integrating new vector databases and embedding services towards our goal of a no/low-code semantic search tool that is easy to configure and deploy. Our first release supported a single ve

bottom of page