Starstruk Logo MarkStarstruk

v3.4.3

DocumindHQ/documindv3.4.3Mar 13, 2026by assafelovic

AI Summary

This is a significant release featuring major fixes and performance improvements, largely driven by contributions from @Br1an67. It introduces new capabilities for tracking research costs and enhancing the FastAPI frontend, alongside various stability fixes and documentation updates.

Key Highlights

  • Major contributions and fixes from @Br1an67 including PyMuPDF timeout improvements and Azure embeddings fallback.
  • New features: Cost tracking per research step, scrape count options, and pondering questions in the FastAPI frontend.
  • Stability enhancements: Import cache invalidation, improved retry handling, and table styles in reports.
  • Documentation updates including fixed Discord links and the addition of an AG2 pipeline diagram.

New Features

  • Track cost per step in research process
  • Add scrape count option to FastAPI frontend
  • Add pondering questions section to FastAPI frontend
  • Add support for connection headers in MCP client/server config

Full Release Notes

Huge release with major fixes and performance improvements. Welcome to all the incredible new contributors and shoutout to @Br1an67 with major contributions!

## What's Changed
* fix: Read all pages in PyMuPDFScraper instead of just first page by @MattBenesch in https://github.com/assafelovic/gpt-researcher/pull/1623
* docs: fix invalid Discord invite link in LLM docs by @Br1an67 in https://github.com/assafelovic/gpt-researcher/pull/1649
* fix: invalidate import caches after runtime package install by @Br1an67 in https://github.com/assafelovic/gpt-researcher/pull/1650
* fix: fallback to OPENAI_API_VERSION for Azure embeddings by @Br1an67 in https://github.com/assafelovic/gpt-researcher/pull/1651
* fix: increase PyMuPDF download timeout and add SSL fallback by @Br1an67 in https://github.com/assafelovic/gpt-researcher/pull/1652
* fix: use all configured retrievers in deep research planning by @Br1an67 in https://github.com/assafelovic/gpt-researcher/pull/1653
* fix: add table styles to FastAPI frontend report container by @Br1an67 in https://github.com/assafelovic/gpt-researcher/pull/1654
* feat: track cost per step in research process by @Br1an67 in https://github.com/assafelovic/gpt-researcher/pull/1655
* feat: add scrape count option to FastAPI frontend by @Br1an67 in https://github.com/assafelovic/gpt-researcher/pull/1656
* feat: add pondering questions section to FastAPI frontend by @Br1an67 in https://github.com/assafelovic/gpt-researcher/pull/1657
* fix: dynamically retrieve agent role if missing in ReportGenerator by @kiwabb in https://github.com/assafelovic/gpt-researcher/pull/1658
* fix docs links and add ag2 pipeline diagram by @qingyun-wu in https://github.com/assafelovic/gpt-researcher/pull/1662
* fix: improve retry handling in create_chat_completion by @jhyz in https://github.com/assafelovic/gpt-researcher/pull/1664
* Add support for connection headers in MCP client/server config by @GeorgelPreput in https://github.com/assafelovic/gpt-researcher/pull/1665

## New Contributors
* @MattBenesch made their first contribution in https://github.com/assafelovic/gpt-researcher/pull/1623
* @Br1an67 made their first contribution in https://github.com/assafelovic/gpt-researcher/pull/1649
* @kiwabb made their first contribution in https://github.com/assafelovic/gpt-researcher/pull/1658
* @qingyun-wu made their first contribution in https://github.com/assafelovic/gpt-researcher/pull/1662
* @jhyz made their first contribution in https://github.com/assafelovic/gpt-researcher/pull/1664

**Full Changelog**: https://github.com/assafelovic/gpt-researcher/compare/v3.4.2...v3.4.3