Releases: langchain-ai/langgraph-studio
Releases · langchain-ai/langgraph-studio
0.0.30
- Add error page & crash ID reporting for better debugging
- Fix issues with
/bin/csh
and/bin/tcsh
shell environments - Fix various crash conditions related to incomplete graph data
- Fix crash when non-string-array value enteres the edit form
- Actually check for presence of docker before returning PATH, support user CLI installation
0.0.29
- Fix streaming errors when
checkpoint_ns
is undefined for old deployments - Check for the API capabilities when loading Studio
- Fix incorrect rendering of errors when returned as object
- Fix editing of pip_config_file within UI changing the
docker_compose_file
as well
0.0.28
- Add support for subgraphs (graph view, see subgraph checkpoint states, replay and edit inner subgraph nodes)
- Show errors in thread view
- Add button to clear logs
- Show progress bar / error when fetching threads
- Reduce flickering of thread view when cancelling runs
0.0.27
0.0.26
- Add CRUD UI for memory store UI
- Add option to pick different editor when opening project (Cursor, PyCharm, custom editor)
- Fix "Open in LangSmith" not redirecting correctly for EU Cloud users
- Fix dark mode bug when editing thread name
- Remove assistants force refresh when switching asisstants
0.0.25
- Add assistant editor, read docs: https://langchain-ai.github.io/langgraph/cloud/how-tos/assistant_versioning/#how-to-version-assistants
- Prevent snapback to previous assistant ID when switching threads
- Remember interrupts when switching assistants
0.0.24
0.0.23
- Adding templates support in Studio. Read more: https://blog.langchain.dev/launching-langgraph-templates/
0.0.22
0.0.21
- Use cache mount for pip cache to speed up Docker builds
- Respect user-provided LangSmith API keys when specified via .env
- Fix broken build when working directory is unknown #105
- Handle message types behind
oneOf
- Fix optional fields being omitted from input editor
- Fix configurable fields resetting after reopening configuration popover #109