July 2025

July 7, 2025
v0.2.19

Client Verbosity

  • Fixed verbose request progress bar in Client.
July 6, 2025
v0.2.18

Simplified Query Client

  • Added a simple method for executing queries while retaining support for customizable queries.
  • Improved bulk request logic.
  • Updated and added demos for both query methods.
  • Updated error messages for failed queries to only print if verbose is set.

Knowledge Graph

  • Added OpenAI-driven knowledge graph extraction with deduplication and improved insertion logic. (apps/kg1.py)
  • Added spaCy-driven knowledge graph extraction for entity/relationship extraction. (apps/kg2.py)
July 2, 2025
v0.2.17

Client Verbosity

  • Progress bars in query execution can now be toggled via the verbose flag (default True).

Providers

  • Added support for OpenAI models (gpt-4o) for knowledge graph extraction.

Knowledge Graph

  • Used OpenAI for entity/edge extraction and improved deduplication and insertion logic.
  • Added FastAPI-based HTTP service for embedding text using HuggingFace transformers in apps/texttovec.py.
  • Added a new example for vector insert operations in examples/vec.py.

June 2025

June 26, 2025
v0.2.16

Providers

  • Added Ollama provider to interface with Ollama for language model tasks.

Knowledge Graph

  • Added a script to use the new Ollama provider to extract entities and relationships from text.
June 23, 2025
v0.2.15

Instance Management

  • Fixed port not updated after deploy.
  • Added auto stop instance on script exit.
  • Removed auto-start instance on client creation.

UV Installation

  • Added support for UV installation.

Knowledge Graph

  • Added a script to take a PDF file as input, extract text from it, and then use a local model to generate embeddings for text chunks.
June 21, 2025
v0.2.11

Instance Management

  • Added instance management via Python script.

May 2025

May 8, 2025
v0.2.1

Initial Release With Official Rag Support

  • This release includes initial support for creating ai rag agents.
  • Supplied are 2 new default queries ragloaddocs and ragsearchdocs with which you can load your chunks of documents and vectors and then vector search for them all in helix-db.
  • Release also contains some bug fixes and more examples + rag demos that can be found in the examples/ directory.