Releases: assafelovic/gpt-researcher
New NextJS front end application and more!
We're excited to officially launch the new and production ready front end application to GPT Researcher leveraging NextJS and Tailwind CSS! This is another important milestone towards the ultimate automated research assistant and big shoutout to @ElishaKay for leading this all the way. In addition, we have some powerful additions by the best community on GIthub including improvements to vector stores, embedding models, usage of env vars with multi agent framework, bug fixes and more! Check the detailed list below.
What's Changed
- [Fix] No module named 'json5' by @gkhngyk in #770
- ✅ CI / CD - run tests on every commit by @ElishaKay in #690
- Update README.md by @mirkancal in #778
- Fix TypeError in custom reports due to missing 'tone' parameter by @Gitzoz in #784
- Improved azure openai config by @danieldekay in #786
- fix: added missing langgraph to requirements.txt by @danieldekay in #785
- ✅ ready for merge: Fullstack fixes by @ElishaKay in #783
- Enhancement: OpenAI Embedding Model Selection by @eitag-uni in #789
- Update prompts.py by @Einengutenmorgen in #797
- multi-agent approach should now use .env file for config of llm by @danieldekay in #795
- Support langchain vectorstores by @hslee16 in #781
- ✅ ready for merge: nextjs to root by @ElishaKay in #800
- multi_agents/requirements.txt was missing loguru by @sdisaacson in #802
New Contributors
- @mirkancal made their first contribution in #778
- @Gitzoz made their first contribution in #784
- @eitag-uni made their first contribution in #789
- @Einengutenmorgen made their first contribution in #797
- @sdisaacson made their first contribution in #802
Full Changelog: v.3.0.5...v3.0.6
Contributors
Assets 2
Performance Boost
This week includes a critical update to all GPTR users - we've focused on optimizations and happy to update that GPTR pip package is now 87% lighter (!!). This is big news to all Docker users and has been a long time request. In addition, we've added support for human in the loop to the multi agent solution. This itself marks a new milestone toward better human-agent experience throughout the research processes. Thank you again to the #1 AI community!
What's Changed
- Stabilize master by @ElishaKay in #755
- Optimized project dependencies by @assafelovic in #756
- minimalistic-deps by @ElishaKay in #757
- made use of azure_openai consistent (and not just in the base model, … by @danieldekay in #759
- WIP: Human in the loop - frontend support by @ElishaKay in #762
New Contributors
- @danieldekay made their first contribution in #759
Full Changelog: v3.0.4...v.3.0.5
Hybrid Research support
I get excited with every release but for a reason! We're now releasing support for hybrid research and multiple retrievers! 🤯
You can now conduct research using both local documents and web search for optimal research results! In addition, you can add multiple web retrievers for even better breadth of online information with any of our supported web retrievers. We've tested this on several research queries and found and increase of 38% in research report quality and correctness!
Thank you again to the #1 community for pushing this forward. This week's special shoutout goes to @0x11c11e for some amazing contributions.
What's Changed
- fix unimport ExaSearch when use Exa by @fei0810 in #722
- Multi Agents - Human in the Loop by @assafelovic in #733
- Correct typo in class name HuggingFaceProvider by @wuloi in #735
- moving the sources box above the logs that follow it by @ElishaKay in #738
- Add Support for Multiple Retrievers in GPTR by @0x11c11e in #726
- chore: Update generate_report function to include tone parameter by @0x11c11e in #741
New Contributors
Full Changelog: v3.0.3...v3.0.4
v3.0.3
Excited to release our latest innovation in multi agent long report collaboration! Thanks to @DandinPower we've made progress in building long reports based on parallel agent work on a given topic with minimal redundancy in report quality. This is another step in building the leading AI research agent. Thanks to the amazing community for further pushing boundaries!
What's Changed
- Async fix by @ruzicka02 in #705
- Add PubMed Central retriever by @0x11c11e in #711
- Feature: Reducing redundancy problems in Detailed Reports by @DandinPower in #713
New Contributors
- @DandinPower made their first contribution in #713
Full Changelog: v3.0.2...v3.0.3
v3.0.2
This week includes improves UX/UI features such as selecting specific research sources, additional retrieval for scholars and the new GPT 4o mini to replace gpt-3.5-turbo. Thank you to the community for this great contributions!
What's Changed
- Update openai GPT models by @adrianhensler in #692
- Enhance File Upload Component with Improved UI and Delete Functionality by @muhammedsaidckr in #703
- Source url research feature added and Fixes report_source and report_type value on multi agents side by @muhammedsaidckr in #700
- Feature/add semantic scholar by @0x11c11e in #687
Full Changelog: v.0.3.1...v3.0.2
v.0.3.1
What's Changed
- Fix ollamaprovider error by @arsaboo in #667
- Fix tone issue related to the default constructor of GPTResearcher by @0x11c11e in #674
- added the unify provider to gpt-researcher by @vedpatwardhan in #661
- Sanitize Directory Names to Ensure Compatibility with Windows by @0x11c11e in #679
- Bugfix: Enable use of ollama and searx without commercial services / Remove openai and tavily dependency (and fixed missing tone errors) by @xdadrm in #681
- Fixes: Default retriever, sanitizing by @ruzicka02 in #683
- Added gptr-logo to frontend/static by @muhammedsaidckr in #682
- Make header an optional argument in TavilyClient by @nikkag in #685
- docs: add Japanese README by @eltociear in #686
- add support for markdown file download from UI by @pkandarpa-cs in #688
New Contributors
- @vedpatwardhan made their first contribution in #661
- @xdadrm made their first contribution in #681
- @ruzicka02 made their first contribution in #683
- @muhammedsaidckr made their first contribution in #682
- @nikkag made their first contribution in #685
- @pkandarpa-cs made their first contribution in #688
Full Changelog: v0.3.0...v.0.3.1
v0.3.0
We're excited to introduce our latest release that includes a completely new UX/UI frontend built with NextJS! This is a huge milestone for the GPT Researcher community, now enabling not only an in depth high quality research agent, but also an optimal research user experience. In addition the release includes some additional performance and stability improvements. Thanks again the amazing community! Check out the new frontend in action:
videoplayback.2.mp4
What's Changed
- Add tone selection feature to research report by @0x11c11e in #654
- Fixed the issue with the generate_report_prompt function to accept the tone parameter by @0x11c11e in #659
- Move blocking operations to threads by @tv2-vonstring in #660
- react frontend 2.0 by @ElishaKay in #523
New Contributors
- @tv2-vonstring made their first contribution in #660
Full Changelog: v0.2.8...v0.3.0
Stability and Quality improvements
This release includes additional search retrievers (Arxiv and more!), and some important stability and quality improvements from improved research quality to UX/UI.
What's Changed
- chore: Update retrievers imports by @0x11c11e in #631
- Correct issue #623 where references from proir runs may remain. by @adrianhensler in #632
- Adding Arxiv Search by @Ayman-Ameen in #543
- Support langchain documents by @hslee16 in #630
New Contributors
- @0x11c11e made their first contribution in #631
- @adrianhensler made their first contribution in #632
- @Ayman-Ameen made their first contribution in #543
- @hslee16 made their first contribution in #630
Full Changelog: v0.2.7...v0.2.8
Stability improvements
What's Changed
- Update introduction.md fixing broken link by @AADaoud in #606
- update on using json_repair from the author of json_repair, thanks! by @pfrench42 in #607
- extract_headers() was throwing an exception when it was examining a l… by @pfrench42 in #610
- Implement more graceful JSON handling on choose_agent in master/actions.py by @AADaoud in #616
New Contributors
Full Changelog: v.0.2.6...v0.2.7
v.0.2.6
This release includes additional stability and performance improvements. In addition, excited to support custom retrievers - Custom retrievers allow you to use any search engine that provides an API to retrieve documents and is widely used for enterprise research tasks. Thanks again to the amazing community for all the contributions!
What's Changed
- change deployment name to model to fix the init of the azureopenai model by @xiepei1 in #594
- use repair_json() on output of sub_queries to stop exceptions on local models by @pfrench42 in #596
- fix: similarity_threshold config by @gkhngyk in #597
- Lazily import langchain-huggingface to avoid obligatory install by @ldorigo in #599
- added implementation for custom retriever by @assafelovic in #602
New Contributors
- @xiepei1 made their first contribution in #594
- @pfrench42 made their first contribution in #596
Full Changelog: v.0.2.5...v.0.2.6
