akhaliq/GLM-5.3-Flash

Capture reasoning_content, set max_tokens, surface empty-response diagnostics

Skip stream chunks with empty choices (fixes IndexError on long answers)

Update workflow.json from canvas

Re-add write-repos scope so owner can save workflow edits

Show friendly sign-in message instead of error when logged out

Bump Gradio to 6.26.0, drop write-repos OAuth scope

Use visitor OAuth token instead of HF_TOKEN env secret

Move workflow app into run.py (Space app_file) and add openai dependency

Use identifier-safe fn bind key to fix node run endpoint

Add GLM-5.3-Flash chat workflow via HF Inference Providers

initial commit

Duplicate from gradio-templates/workflow