Repository navigation

All pull requests

Labels (55)Milestones (0)New pull request

Open

(952)

Closed

(10,029)

docs: add Dash0 to the observability integrations
#24114CodingFabian opened Nov 10, 20264 tasks done
fix(core): HierarchicalNodeParser async parsing returns documents unsplit
#2411367-luck opened Nov 10, 20266 of 8 tasks
fix(duckdb): compare numeric metadata filters as DOUBLE
#24112HhH666-ux opened Nov 10, 2026
fix(duckdb): handle empty metadata membership filters
#24111jjhong011001 opened Nov 10, 20261 task done
docs: add local OpenAILike recipe with explicit capabilities
#24110Mansi2221 opened Nov 10, 2026
1
fix(core): await achat in MultiModalRelevancyEvaluator.aevaluate and preserve EvaluationResult fields
#24109deep1blue opened Nov 10, 20268 tasks done
docs: list the LinkDigest reader under community Data Loaders
#24108jcaiagent7143-ui opened Nov 10, 20266 of 8 tasks
fix(core): update document references when nodes change sources
#24106remote-controlled-man opened Nov 9, 20263 tasks done
1
fix(core): apply embedding batch rate limits only to cache misses
#24105remote-controlled-man opened Nov 9, 20264 tasks done
1
fix(core): do not execute tool calls from a reply the provider marked truncated
#24103simin-yuan opened Nov 9, 2026
perf(core): cache embeddings and spaCy docs in SemanticDoubleMergingSplitterNodeParser
#24101deep1blue opened Nov 9, 2026
state prompt formatting rewrites the user's message in agent memory
#24099goingforstudying-ctrl opened Nov 9, 2026
fix(core): return ints in extract_numbers_given_response and reject 0 index in tree index
#24097deep1blue opened Nov 9, 20263 tasks done
2
feat(assemblyai-reader): add lazy loading, transcript_id support, and new formats
#24096pylrn opened Nov 9, 202613 of 16 tasks
docs: add physics/bounds validation Workflow example
#24094ZYZ666-RGB opened Nov 9, 20261 of 2 tasks
6
fix(core): prevent stale evaluation reuse in FeedbackQueryTransformation
#240931105216375-alt opened Nov 9, 20262 tasks done
fix(core): isolate tool results across agent runs after failure
#24092remote-controlled-man opened Nov 9, 20267 of 14 tasks
1
docs: add Vietnamese PDF ingestion example using the current Document API (#23708)
#24090wangkangmao opened Nov 8, 2026
1
fix(core): preserve query_transformer and callback_manager across retries in RetryQueryEngine
#24089deep1blue opened Nov 8, 2026
2
fix(core): retain filesystem when restoring graph store
#24088Ricardo-M-L opened Nov 8, 2026Contributor3 tasks done
fix(core): preserve KV store snapshot on serialization failure
#24087Ricardo-M-L opened Nov 8, 2026Contributor3 tasks done
fix(core): scope property graph node deduplication by identity
#24084remote-controlled-man opened Nov 8, 20264 tasks done
1
fix(core): avoid defaultdict mutation and plain-dict KeyError in LlamaDebugHandler
#24083deep1blue opened Nov 8, 20263 tasks done
fix(readers-web): keep RSS user agents local to each reader
#24081wToTw123 opened Nov 8, 20268 of 15 tasks
perf(evaluation): reuse QA query engine and batch async QA in DatasetGenerator
#24080deep1blue opened Nov 8, 20269 tasks done

Navigation Menu

Repository navigation

All pull requests

docs: add Dash0 to the observability integrations

fix(core): HierarchicalNodeParser async parsing returns documents unsplit

fix(duckdb): compare numeric metadata filters as DOUBLE

fix(duckdb): handle empty metadata membership filters

docs: add local OpenAILike recipe with explicit capabilities

fix(core): await achat in MultiModalRelevancyEvaluator.aevaluate and preserve EvaluationResult fields

docs: list the LinkDigest reader under community Data Loaders

fix(core): update document references when nodes change sources

fix(core): apply embedding batch rate limits only to cache misses

fix(core): do not execute tool calls from a reply the provider marked truncated

perf(core): cache embeddings and spaCy docs in SemanticDoubleMergingSplitterNodeParser

state prompt formatting rewrites the user's message in agent memory

fix(core): return ints in extract_numbers_given_response and reject 0 index in tree index

feat(assemblyai-reader): add lazy loading, transcript_id support, and new formats

docs: add physics/bounds validation Workflow example

fix(core): prevent stale evaluation reuse in FeedbackQueryTransformation

fix(core): isolate tool results across agent runs after failure

docs: add Vietnamese PDF ingestion example using the current Document API (#23708)

fix(core): preserve query_transformer and callback_manager across retries in RetryQueryEngine

fix(core): retain filesystem when restoring graph store

fix(core): preserve KV store snapshot on serialization failure

fix(core): scope property graph node deduplication by identity

fix(core): avoid `defaultdict` mutation and plain-dict `KeyError` in `LlamaDebugHandler`

fix(readers-web): keep RSS user agents local to each reader

perf(evaluation): reuse QA query engine and batch async QA in DatasetGenerator

Pull requests