We’re excited to launch the SWE-agent web UI! Specify a bug, press start and watch SWE-agent do the magic.
- Run SWE-agent in the cloud using GitHub Codespaces
- Add GPT4-turbo model by @zgrannan in princeton-nlp#252
- feat: Amazon Bedrock support (Claude models) by @JGalego in princeton-nlp#207
- Better error handling for --open_pr by @klieret in princeton-nlp#239
- Fixed a potential error by @DanjieTang in princeton-nlp#242
- fix: TARGETARCH not set on some OS/docker setups by @mspronesti in princeton-nlp#249
- Pass Python version to get_environment_yml by @waterson in princeton-nlp#271
- Fix Together model validation error by @mikanfactory in princeton-nlp#236
- Doc: Avoid invalid github token by @klieret in princeton-nlp#292
- Allow to run on local repos (new flag:
--repo_path
) in princeton-nlp#193 - Patch files are now saved separately to a patch directory in princeton-nlp#126
- Allow to supply custom installation commands when running on gh issues or locally (
--environment_setup
) in princeton-nlp#153 - Allow to specify openapi base url in
keys.cfg
in princeton-nlp#118
- Improve error handling of docker issues in princeton-nlp#165
- Make github token fully optional in princeton-nlp#189
- Fix opening PR from fork in princeton-nlp#229
- Fix: Choosing TogetherAI models in princeton-nlp#130