aalekh-sarvam
7cd051d7f7
fix(memory): use sqlite-vec KNN for searchVector (190x speedup)
...
Replace full-table scan via vec_distance_cosine() + ORDER BY LIMIT with
sqlite-vec's native MATCH + k = ? KNN operator. Keep vec_distance_cosine()
in the SELECT so score = 1 - dist preserves the existing cosine [0,1]
semantics the downstream merge pipeline depends on.
Fixes #69666 .
Benchmark on 10,827 chunks, 4096-dim embeddings:
- Before (full scan): ~8490 ms/query
- After (KNN + join): ~50 ms/query
No behavioral changes: returned ids and ordering are identical to the
previous query on all tested queries. The LIMIT ? binding is replaced by
k = ? which caps sqlite-vec's candidate set to the same count.
2026-04-23 04:01:47 +01:00
Bek
dfe0e49c8a
fix(qmd): Dedup in-flight manager creation so only one full QMD manager arms per agent/config at a time, eliminating the concurrent exportSessions() collisions that triggered path changed during write errors ( #65226 )
...
Fixes concurrent manager creation races that caused SafeOpenErrors during session export.
Deduplicates in-flight manager creation so only one full QMD manager arms per agent/config at a time, eliminating the concurrent exportSessions() collisions that triggered path changed during write errors
Resolves and snapshots runtime inputs before cache reuse, replacing stale managers atomically when workspace/config changes, and aborting queued export work promptly on close()
2026-04-21 18:22:21 -04:00
Peter Steinberger
512dc4f2b1
test: share memory session search setup
2026-04-20 21:14:46 +01:00
Peter Steinberger
b3a97df754
refactor: cache reply and visibility runtimes
2026-04-18 20:54:30 +01:00
Peter Steinberger
df525b90f2
chore(lint): enable unnecessary type parameter rule
2026-04-18 18:31:13 +01:00
Ayaan Zaidi
26f7198eda
fix(memory-core): preserve vector dims on readonly recovery
2026-04-17 11:22:56 +05:30
Rubén Cuevas
7b0e950e09
fix: dedupe degraded sqlite-vec warnings ( #67898 ) (thanks @rubencu)
...
* Agents: dedupe bootstrap truncation warnings
* Memory: dedupe sqlite-vec degradation warnings
* Memory: align degraded vector warning
* test(memory-core): remove stale vector warning arg
* fix(memory-core): reset degraded warning on vector reset
* fix(memory-core): preserve warning latch across reindex rollback
* fix: dedupe degraded sqlite-vec warnings (#67898 ) (thanks @rubencu)
---------
Co-authored-by: Ayaan Zaidi <hi@obviy.us >
2026-04-17 11:09:14 +05:30
Peter Steinberger
77e6e4cf87
refactor: move memory embeddings into provider plugins
2026-04-17 02:57:18 +01:00
Peter Steinberger
4efd3c3d74
test: harden beta release gates
2026-04-15 19:28:49 +01:00
Tak Hoffman
4f00b76925
fix(context-window): Tighten context limits and bound memory excerpts ( #67277 )
...
* Tighten context limits and bound memory excerpts
* Align startup context defaults in config docs
* Align qmd memory_get bounds with shared limits
* Preserve qmd partial memory reads
* Fix shared memory read type import
* Add changelog entry for context bounds
2026-04-15 13:06:02 -05:00
Agustin Rivera
37d5971db3
Align QMD memory reads with canonical memory paths ( #66026 )
...
* fix(memory): align qmd read paths
Co-authored-by: zsx <git@zsxsoft.com >
* fix(memory): add qmd exact-path read fast path
* fix(memory): tighten qmd read-path guards
* changelog: note QMD memory_get canonical-path restriction (#66026 )
---------
Co-authored-by: zsx <git@zsxsoft.com >
Co-authored-by: Devin Robison <drobison@nvidia.com >
2026-04-14 09:58:27 -06:00
Vincent Koc
37f449d7e1
fix(memory): restore ollama embedding adapter ( #66269 )
...
* fix(memory): restore ollama embedding adapter
* Update CHANGELOG.md
2026-04-14 09:02:31 +01:00
Mariano
3d06d90e83
fix(memory): unify default root memory handling ( #66141 )
...
* fix(memory): unify default root memory handling
* test(memory): align legacy migration expectation
* docs(changelog): tag qmd root-memory fix
* docs(changelog): append qmd root-memory entry
* docs(changelog): dedupe qmd root-memory entry
* docs(changelog): attribute qmd root-memory fix
---------
Co-authored-by: mbelinky <mbelinky@users.noreply.github.com >
2026-04-13 23:59:57 +02:00
Peter Steinberger
645c7b1897
fix: harden qmd service startup
2026-04-13 10:58:46 +01:00
Rugved Somwanshi
0cfb83edfa
feat: LM Studio Integration ( #53248 )
...
* Feat: LM Studio Integration
* Format
* Support usage in streaming true
Fix token count
* Add custom window check
* Drop max tokens fallback
* tweak docs
Update generated
* Avoid error if stale header does not resolve
* Fix test
* Fix test
* Fix rebase issues
Trim code
* Fix tests
Drop keyless
Fixes
* Fix linter issues in tests
* Update generated artifacts
* Do not have fatal header resoltuion for discovery
* Do the same for API key as well
* fix: honor lmstudio preload runtime auth
* fix: clear stale lmstudio header auth
* fix: lazy-load lmstudio runtime facade
* fix: preserve lmstudio shared synthetic auth
* fix: clear stale lmstudio header auth in discovery
* fix: prefer lmstudio header auth for discovery
* fix: honor lmstudio header auth in warmup paths
* fix: clear stale lmstudio profile auth
* fix: ignore lmstudio env auth on header migration
* fix: use local lmstudio setup seam
* fix: resolve lmstudio rebase fallout
---------
Co-authored-by: Frank Yang <frank.ekn@gmail.com >
2026-04-13 15:22:44 +08:00
Peter Steinberger
e76c2812b7
style: apply oxfmt
2026-04-12 11:28:43 -07:00
Vincent Koc
d4fb7d893d
fix(ci): repair main tsgo regressions
2026-04-12 19:14:00 +01:00
jasonxargs-boop
2204753b62
fix(memory-core): fix macOS chokidar glob issue by watching memory dir directly ( #64711 )
...
* fix(memory-core): fix macOS chokidar glob issue by watching memory dir directly
* fix(memory-core): ignore non-markdown memory watch churn
* fix(memory-core): allow multimodal watch events
* test(memory-core): type watcher ignore callback
---------
Co-authored-by: Vincent Koc <vincentkoc@ieee.org >
2026-04-12 18:53:20 +01:00
Peter Steinberger
15b86ac6d0
fix: narrow qmd defaults and clawblocker memory
2026-04-12 18:52:06 +01:00
Anonymous Amit
42590106ab
improve memory fallback lexical ranking ( #65395 )
...
* improve memory fallback lexical ranking
* use neutral lexical fallback fixtures
* fix(memory-core): keep lexical boosts out of hybrid search
---------
Co-authored-by: Vincent Koc <vincentkoc@ieee.org >
2026-04-12 18:36:28 +01:00
Tak Hoffman
885209ed03
feat: default active memory QMD recall to search ( #65068 )
...
* feat(active-memory): default QMD recall to search
* feat(active-memory): surface search debug telemetry
* fix(active-memory): avoid forking qmd managers
2026-04-11 20:56:21 -05:00
Peter Steinberger
d515009c53
fix(ci): stabilize auto-reply CI tests
2026-04-11 04:09:10 +01:00
Peter Steinberger
3e80bd33e4
refactor: simplify extension conversions
2026-04-11 01:37:23 +01:00
Peter Steinberger
d44cd0d452
style: apply oxformat cleanup
2026-04-11 01:17:51 +01:00
Peter Steinberger
ebfd468ee0
refactor: simplify typed conversions
2026-04-11 01:01:30 +01:00
Peter Steinberger
a73dc477d7
style: apply oxfmt cleanup
2026-04-10 23:09:37 +01:00
Peter Steinberger
d236cb4680
chore: enable redundant type constituent checks
2026-04-10 21:23:40 +01:00
Peter Steinberger
f6ed276f51
style: apply updated formatter output
2026-04-10 19:17:39 +01:00
Peter Steinberger
7897fb9c84
refactor: dedupe locale lowercase helpers
2026-04-07 22:57:52 +01:00
Peter Steinberger
dffa88f396
refactor: dedupe memory lowercase helpers
2026-04-07 20:58:01 +01:00
Peter Steinberger
eb9ce9482c
refactor: dedupe memory lowercase helpers
2026-04-07 20:57:04 +01:00
Peter Steinberger
a00b01f5ed
fix: harden complex qa suite scenarios
2026-04-07 20:35:39 +01:00
Peter Steinberger
9314bb7180
refactor: dedupe extension lowercase helpers
2026-04-07 15:12:32 +01:00
Peter Steinberger
e169fcd263
refactor: dedupe qa and diff error formatting
2026-04-07 05:06:54 +01:00
Peter Steinberger
3417dbabf4
refactor: dedupe memory-core error formatting
2026-04-07 02:03:34 +01:00
Peter Steinberger
3a1ca98e53
perf: extract memory multimodal indexing policy
2026-04-07 00:17:08 +01:00
Peter Steinberger
d2a03eca1a
perf: extract memory session sync state helpers
2026-04-07 00:17:08 +01:00
Vincent Koc
32eff914c6
fix(memory-core): narrow qmd and artifact dir typing
2026-04-06 23:09:46 +01:00
Peter Steinberger
e44a995e83
test: trim qmd manager fixture setup
2026-04-06 22:49:36 +01:00
Peter Steinberger
b1905c1423
refactor: dedupe qmd manager record helper
2026-04-06 22:44:14 +01:00
Peter Steinberger
e88c39b0a1
refactor: dedupe memory-core error formatting
2026-04-06 22:44:14 +01:00
Vincent Koc
e92c2b63f9
fix(memory-core): align embedding cache db typing
2026-04-06 22:16:12 +01:00
Peter Steinberger
4ae1599ea5
perf: extract memory adapter registration helper
2026-04-06 22:04:23 +01:00
Peter Steinberger
d806682f78
perf: extract memory embedding state helpers
2026-04-06 22:04:23 +01:00
Vincent Koc
b96589b1fc
fix(memory-core): align vector write db typing
2026-04-06 21:54:32 +01:00
Peter Steinberger
c9f288ceaf
perf: extract memory atomic reindex helpers
2026-04-06 21:28:29 +01:00
Peter Steinberger
6b6c95b443
perf: extract memory sqlite write helpers
2026-04-06 21:28:29 +01:00
Peter Steinberger
ca27d932b4
perf: extract memory search preflight helpers
2026-04-06 21:28:29 +01:00
Vincent Koc
ca26489fe8
fix(memory-core): repair sync helper typing drift
2026-04-06 21:11:06 +01:00
Peter Steinberger
7c256bfdf4
test: trim memory manager test startup
2026-04-06 20:52:08 +01:00