Ad
Haystack Releases
AI and NLP Search Engine Updates
Source status
HealthyThis source is refreshing normally and has recorded updates.
Last checked 13 Oct 2026, 19:47 GMT+2
Latest updates
Summary generated 10 Oct 2026.
- SentenceWindowRetriever now queries the Document Store once per run or run_async call instead of once per retrieved document, reducing latency.
- SummarizationCompactor was added as an experimental compaction strategy for CompactionHook, replacing older turns and Agent steps with LLM-generated summaries.
- InMemoryDocumentStore.bm25_retrieval and InMemoryBM25Retriever now return only documents that contain at least one query term.
Tags
ainlpsearch
Activity Score
79%
Great
Based on recency, cadence, and completeness
83%
Recency
50%
Cadence
97%
Complete
100%
Health
- Adapter
- atom
- Vendor
- deepset
API keys are available
Create a key, follow this source in the app, then request entries for your selected sources. Free covers 10 followed sources; Pro includes unlimited catalog access with REST and local MCP.
GET /api/v1/entries?sources=haystack.releasesAd
Related Sources
Loading source
Loading source