Start with install notes or jump straight into the API.

Overview

Roadmap & Changelog

Up Next

  • ๐Ÿงท Count and scan primitives โ€” filter-count mode, snapshot truncation removal, route renames (RFC 0019, #67)
  • ๐Ÿš‘ Indexing failure-mode E2E runbook โ€” Aerospike stop-writes + Postgres pressure (#55)
  • ๐Ÿงฌ Embedding UDF writeback via row re-upsert (#52)
  • ๐ŸŒฑ Namespace init UDF for first-time embed population
  • ๐Ÿšฆ Snapshot-aware ready signal โ€” layer.is_stable honors UDF state (#46)
  • ๐ŸŽจ Full dashboard redesign โ€” 6-tab layout from the prototype (#11)

Later

  • ๐Ÿ” RBAC: scoped API keys and entitlements as a Layer primitive (#8)
  • โ™ป๏ธ Soft delete with TTL + restore (#7)
  • ๐Ÿชข Hybrid text search via RRF (#18)
  • โŒจ๏ธ Typeahead via Turbopuffer regex index (#19)
  • ๐Ÿ•ฐ๏ธ Temporal queries โ€” as_of selector for /query, /scans, /fetch, and /snapshots (RFC 0020, #68)
  • ๐ŸŒฟ copy_from_with_filter โ€” time travel + subset branching (#20)
  • ๐Ÿ‡ Exact kNN result cache keyed by consistency watermark (#21)
  • ๐Ÿงช A/B variant indexes with operator-controlled rollout (#23)
  • ๐Ÿฆš Per-query observability with LLM-judged Tail Quality (#41)
  • ๐ŸŽž๏ธ Pipeline crash recovery via source replay + deterministic IDs (#43)
  • โ˜ ๏ธ Paginated UDF dead-letter list (#44)
  • ๐Ÿ—๏ธ Narrow cluster topology defaults (#45)
  • ๐Ÿ“ฃ Write amplification baselines (#15)
  • ๐Ÿ“ฎ layer push โ€” Python UDF dev experience via Depot (#64)

0.1 Release (UAT)

API hardening

  • ๐Ÿงฉ Scaling CRD consolidation โ€” Pipeline, UDF, InfraRules (RFC 0012)
  • ๐ŸŽ›๏ธ Index CRD redesign (RFC 0013)
  • ๐Ÿ“ธ Snapshot scan naming conventions (RFC 0014)
  • ๐Ÿงน Remove unused APIs

Lifecycle and operability

Surfaces

  • ๐ŸชŸ Dashboard MVP โ€” basic CRD management and observability
  • ๐Ÿš layer CLI MVP โ€” kube-style commands for Layer
  • ๐Ÿ Official Python SDK
  • ๐ŸŽฏ Strongly consistent queries during heavy writes via _hevlayer_upserted_at
  • ๐Ÿงฎ Result count over FTS/vector queries via scatter/gather
  • ๐Ÿ“œ Precomputed facet listings in snapshots
  • ๐Ÿช™ Precomputed facet counts in snapshots
  • ๐Ÿชƒ Scans for filter IDs and filter counts not available in a snapshot
  • ๐Ÿ†” Search by id via document-cached vector
  • ๐Ÿ“ฐ Search history saved to S3
  • ๐Ÿ—‚๏ธ Enhanced namespace metadata