Ai

GPT Researcher v3.4.4 Improves Retrieval, Parsing and Provider Support

GPT Researcher v3.4.4 Improves Retrieval, Parsing and Provider Support

GPT Researcher v3.4.4 is a focused maintenance and capability release that improves the reliability of research workflows while adding new model and retrieval options. This update introduces native MiniMax support for LLM and embeddings, adds an X/Twitter search retriever, and resolves a series of issues affecting scraping, parsing, PubMed retrieval, websocket handling, PDF reports, and streaming output.

What Changed

The most notable feature addition in v3.4.4 is native support for MiniMax as both an LLM and embedding provider. This expands deployment flexibility for teams evaluating model backends and gives developers another built-in option for powering research pipelines.

The release also adds a new Xquik X/Twitter search retriever, broadening the range of sources GPT Researcher can query during information gathering. For users building workflows around social content discovery and trend monitoring, this is a meaningful expansion of retrieval coverage.

On the stability side, the release includes fixes for a reference error in code, potential data parsing issues in web scraping, and context normalization for both dictionary and string formats. These changes should reduce edge-case failures and improve consistency when handling mixed input structures.

Several targeted bug fixes also address missing os imports, PubMed Central retrievers returning no results, PDF report issues, and a separate issue tracked as #1712. Streaming behavior was improved by skipping empty LLM output chunks and flushing stdout more cleanly, which should help produce more reliable live output during long-running research tasks.

Finally, dependency updates are included as part of the release, helping keep the project current and reducing the risk of compatibility drift.

Why It Matters

For AI teams using GPT Researcher in production or internal tooling, v3.4.4 is important less as a headline-grabbing release and more as a practical quality upgrade. It strengthens the core pipeline where research systems often break first: source retrieval, parsing, output generation, and provider integration.

The addition of MiniMax support makes the platform more adaptable in a fast-moving model ecosystem, while the new X/Twitter retriever improves access to timely public content. At the same time, the numerous bug fixes suggest a release aimed at reducing friction in real-world usage, especially in workflows that depend on web scraping, academic retrieval, streaming responses, and generated reports.

Overall, GPT Researcher v3.4.4 looks like a solid incremental release that improves robustness and broadens the platform's usefulness for developers building AI-assisted research and information synthesis products.

Official Source: https://github.com/assafelovic/gpt-researcher/releases/tag/v3.4.4