Repository Issue Activity (beta)

#1193 Latest version of LiteLLM is broken with Python 3.10+

#1225 [BUG] ServiceTask requests have no timeout and can hang a workflow indefinitely

#1161 `mcpadapt` not working with latest version of `mcp`

#1221 Update vectors package to standardize fallback logic

#1211 Add ONNX vectors backend

#1209 Bash tool hangs forever on an allowed command that reads stdin

#741 Feature request : LLM Integration for Knowledge Graph Enhancement

#1173 Create default LEMUR model

#1199 Idea: portable regression-eval suite (EvalPort) for RAG pipeline outputs

#1203 Faiss default nprobe derives from corpus count, not the trained nlist

#1204 [bug] ANN backends crash saving safetensors storage since 9.9.0 (regression)

#1197 WordVectors.index() accepts a checkpoint parameter but never uses it

#1185 Add max pooling

#1189 Similarity/CrossEncoder(crossencode=True) ranks candidates by inconsistent labels with multi-class models (e.g. NLI cross-encoders)

#1174 Agent disabled and smolagents wrongly blamed when installing only the "agent" extra

#1182 OpenAI chat route sends the oldest message to non-LLM backends

#1187 PoolingFactory.method raises KeyError on a pooling config without pooling_mode_mean_tokens

#1190 LEMUR training/indexing mismatch with batch-scope centering

#1191 Show your engineering momentum: a one-line badge for your README

#1024 Feature request: Add LEMUR: Learned Multi-Vector Retrieval

#1181 Add `safeopen` validation to `retrieve` task output

#1180 Refactor late encoding logic to cleanup logic paths

#1167 Convert MUVERA algorithm from NumPy to Torch

#1176 Latest version of `docling` causing Windows build errors

#1169 Update LEMUR to follow `txtai` coding conventions