What's Changed

• docs: clarify filter prefix requirements by @haosenwang1018 in #1352
• docs(common): add package readme metadata by @haosenwang1018 in #1338
• docs(client): prefer released pip install instructions by @haosenwang1018 in #1335
• fix(ts-client): include readme in published package by @haosenwang1018 in #1340
• feat(client+langgraph): raw filter strings and EpisodeType normalization by @sscargal in #1403
• Build(deps-dev): Bump ty from 0.0.32 to 0.0.35 by @dependabot[bot] in #1401
• Build(deps): Bump the npm_and_yarn group across 2 directories with 6 updates by @dependabot[bot] in #1393
• Wire EventMemory as backend for LongTermMemory by @edwinyyyu in #1395
• Add Agent skill by @malatewang in #1397
• docs: regenerate openapi.json by @github-actions[bot] in #1404
• docs: clarify filter field prefix requirements by @haosenwang1018 in #1333
• Build(deps): Bump the uv group across 1 directory with 2 updates by @dependabot[bot] in #1392
• fix: resolve ty type-check issues in packages/ by @edwinyyyu in #1405
• Build(deps): Bump pgvector from 0.4.1 to 0.4.2 by @dependabot[bot] in #1398
• Build(deps): Bump langchain-aws from 1.2.1 to 1.4.6 by @dependabot[bot] in #1400
• Build(deps): Bump uvicorn from 0.40.0 to 0.46.0 by @dependabot[bot] in #1399

New Contributors

• @haosenwang1018 made their first contribution in #1352

Full Changelog: v0.3.8...v0.3.9

What's Changed

• Embedded VectorStore Implementation(s) by @edwinyyyu in #1292
• docs: regenerate openapi.json by @github-actions[bot] in #1385
• Fix Installation Test CI by @edwinyyyu in #1389
• update legacy LoCoMo eval scripts to current MemMachine API by @1saac-k in #1272
• New long-term episodic memory (named EventMemory to differentiate) based on VectorStore and (new) SegmentStore by @edwinyyyu in #1205
• feat: support separate LLM models for agent, answer, and judge in ben… by @cudaMancpy in #1376
• Segmenter and Deriver should be async by @edwinyyyu in #1394
• Don't ensure ASCII in LLM-visible JSON strings by @edwinyyyu in #1378

New Contributors

• @cudaMancpy made their first contribution in #1376

Full Changelog: v0.3.7...v0.3.8

What's Changed

• Add strands-memmachine integration for Strands Agents by @Viranshu-30 in #1330
• fix(ts-client): Auto-select fetch adapter behind HTTPS proxies (#1269) by @alighazi288 in #1369
• Build(deps-dev): Bump ty from 0.0.24 to 0.0.32 by @dependabot[bot] in #1354
• Build(deps): Bump the npm_and_yarn group across 1 directory with 2 updates by @dependabot[bot] in #1368
• Improve vector store API again by @edwinyyyu in #1379
• Add SQL filter compilation for properties JSON by @edwinyyyu in #1291
• Implement QdrantVectorStore by @edwinyyyu in #1086
• Build(deps-dev): Bump complexipy from 5.3.0 to 5.4.0 by @dependabot[bot] in #1372
• Support batch size in embedder by @malatewang in #884
• Build(deps-dev): Bump ruff from 0.15.11 to 0.15.12 by @dependabot[bot] in #1373
• Delete session asynchronously by @malatewang in #1117
• Decrease frequency of macOS Intel CI timeout by @edwinyyyu in #1382

New Contributors

• @Viranshu-30 made their first contribution in #1330
• @alighazi288 made their first contribution in #1369

Full Changelog: v0.3.6...v0.3.7

What's Changed

• docs: regenerate openapi.json by @github-actions[bot] in #1351
• Build(deps): Bump the npm_and_yarn group across 2 directories with 5 updates by @dependabot[bot] in #1350
• Build(deps): Bump the npm_and_yarn group across 1 directory with 3 updates by @dependabot[bot] in #1353
• Build(deps-dev): Bump ruff from 0.15.10 to 0.15.11 by @dependabot[bot] in #1355
• Build(deps-dev): Bump complexipy from 5.2.0 to 5.3.0 by @dependabot[bot] in #1356
• Potential fix for code scanning alert no. 46: Prototype-polluting assignment by @sscargal in #1357
• add retry logic for LLM judge and semantic memory consolidation by @sscargal in #1361
• feat: add delete RUN_TYPE to retrieval-agent benchmarks by @sscargal in #1363
• fix(server): preserve original LLM error in ExternalServiceAPIError by @sscargal in #1362
• chore: add workflow to auto-close stale issues and PRs by @sscargal in #1358
• security: bump vulnerable dependencies (Dependabot batch) by @sscargal in #1366
• Increase complete integration test timeout by @edwinyyyu in #1367
• fix: use correct Neo4j config field name in sample configs by @sscargal in #1360

Full Changelog: v0.3.5...v0.3.6

What's Changed

• Build(deps-dev): Bump pytest from 9.0.2 to 9.0.3 by @dependabot[bot] in #1321
• fix(embedder): always chunk text before cluster_texts() when max_input t_length is None by @sscargal in #1328
• Build(deps-dev): Bump ruff from 0.15.9 to 0.15.10 by @dependabot[bot] in #1320
• fix(openclaw): prevent autoRecall cross-session memory leakage (#1318) by @sscargal in #1322
• chore: bump Node 20 → 24 in GitHub Actions workflows by @sscargal in #1323
• feat: enhance memmachine-compose.sh first-run setup wizard by @1saac-k in #1273
• feat(ts-client): support batch memory delete by @yaleMemVerge in #1266
• Build(deps-dev): Update testcontainers[neo4j,postgres] requirement from >=4.13.1 to >=4.14.2 by @dependabot[bot] in #1319
• Add LLM-friendly formatting helpers to clients by @edwinyyyu in #1311

New Contributors

• @1saac-k made their first contribution in #1273

Full Changelog: v0.3.4...v0.3.5

What's Changed

• Security/fix archive path traversal by @o-love in #1287
• Docs: Updating Quickstart for missing backticks and removing the Google ADK Integration Guide by @SarahScargall in #1294
• Improve vector store API by @edwinyyyu in #1201
• Define vector similarity score direction and range by @edwinyyyu in #1296
• docs: sync platform openapi.json by @github-actions[bot] in #1282
• Rename Collection --> VectorStoreCollection by @edwinyyyu in #1301
• text-embedding-3-small special token and intermittent failure fix by @edwinyyyu in #1293
• Index uid and timestamp properties by @edwinyyyu in #1000
• Get config from collection by @edwinyyyu in #1302
• Disable store in openAI API by @malatewang in #1275
• Upgrade openclaw plugin by @malatewang in #1267
• feat: support base_url for Cohere rerankers by @hyeongtakji in #1262
• Fix deadlock due to non-reentrant lock pool by @edwinyyyu in #1312
• refactor: modernize storage interface types by @o-love in #1239
• Support Stream chat completion response by @malatewang in #1276
• fix: NebulaGraph backend - enable concurrent operations and fix datet… by @wenhaocs in #1314
• Build(deps-dev): Bump ruff from 0.15.7 to 0.15.9 by @dependabot[bot] in #1300
• Bump the npm_and_yarn group across 1 directory with 4 updates by @dependabot[bot] in #1290
• Build(deps-dev): Bump handlebars from 4.7.8 to 4.7.9 in /packages/ts-client in the npm_and_yarn group across 1 directory by @dependabot[bot] in #1295
• docs: regenerate openapi.json by @github-actions[bot] in #1315
• feat: add cluster engine for semantic message grouping by @o-love in #1240
• Bump the npm_and_yarn group across 1 directory with 5 updates by @dependabot[bot] in #1284

New Contributors

• @hyeongtakji made their first contribution in #1262

Full Changelog: v0.3.3...v0.3.4

What's Changed

• docs: sync platform openapi.json by @github-actions[bot] in #1237
• Bump authlib from 1.6.7 to 1.6.9 in the uv group across 1 directory by @dependabot[bot] in #1234
• [GH-1147] Fix null byte crash in semantic memory ingestion by @o-love in #1165
• [GH-1146] Cap features sent to LLM during semantic ingestion update by @o-love in #1163
• Handle context window size when creating summary in short term memory by @malatewang in #1254
• fix: replace O(n²) ingestion polling query with GROUP BY/HAVING by @xiongzubiao in #1253
• Docs: Update Nebula Integration Guide and Navigation by @SarahScargall in #1260
• Make benchmark agent LLM configurable through config file by @Tianyang-Zhang in #1217
• Retry OpenAI API on internal server error by @edwinyyyu in #1226
• docs: sync platform openapi.json by @github-actions[bot] in #1250
• Bump ruff from 0.15.5 to 0.15.7 by @dependabot[bot] in #1256
• Bump ty from 0.0.23 to 0.0.24 by @dependabot[bot] in #1257
• Bump pytest-cov from 7.0.0 to 7.1.0 by @dependabot[bot] in #1255
• fix: stop semantic ingestion retry loop on context_length_exceeded by @o-love in #1259

Full Changelog: v0.3.2...v0.3.3

What's Changed

• Strip the leading 'v' from VERSION_INPUT by @sscargal in #1222
• Add LongMemEval evaluation scripts by @edwinyyyu in #1216
• Docs: Helm Chart Cloud Deployment Guide by @SarahScargall in #1224
• Add knowledge update score to lme_generate.py by @edwinyyyu in #1223
• Bump orjson from 3.11.5 to 3.11.6 in the uv group across 1 directory by @dependabot[bot] in #1218
• feat: expose connection pool lifetime and timeout settings for Neo4j and SQLAlchemy by @xiongzubiao in #1228
• Bump ty from 0.0.21 to 0.0.23 by @dependabot[bot] in #1232

Full Changelog: v0.3.1...v0.3.2

What's Changed

• Specify the 'main' branch when checking out code. by @sscargal in #1182
• Correct the working-directory by @sscargal in #1185
• Add workflow_dispatch trigger to OpenClaw plugin publish workflow by @Copilot in #1186
• Fix/openclaw plugin publish by @sscargal in #1187
• update evaluation utils to work with memmachine v0.3 #1196 by @tomw-mv in #1197
• support pvc with standard storageclass by @jgong in #1194
• Auto ts client versioning from git tag on publish by @yaleMemVerge in #1195
• Batch vectors to enable more efficient queries by @edwinyyyu in #1200
• Support new openclaw plugin request by @malatewang in #1191
• Automate docs/openapi.json by @sscargal in #1190
• docs: sync platform openapi.json by @github-actions[bot] in #1183
• docs: regenerate openapi.json by @github-actions[bot] in #1206
• Package Ecosystem Update to address upstream Security fixes by @sscargal in #1198
• Docs: Updating Examples Documentation to reflect V2 Agents by @SarahScargall in #1209
• Bump authlib from 1.6.6 to 1.6.7 in the uv group across 1 directory by @dependabot[bot] in #1208
• Bump ruff from 0.15.1 to 0.15.5 by @dependabot[bot] in #1204
• Bump complexipy from 5.0.0 to 5.2.0 by @dependabot[bot] in #1202
• Docs: Updating Configuration section for recent Semantic Memory Additions and Reorganization by @SarahScargall in #1215
• Docs: Adding new OpenClaw Integration Guide by @SarahScargall in #1207
• [GH-1160] Fix consolidation pipeline silently destroying user-defined tags by @o-love in #1164
• Bump ty from 0.0.20 to 0.0.21 by @dependabot[bot] in #1203
• Docs: Friendly rewrite of Retrieval Agent Architecture page to include Intelligent Orchestration image, match language cohesion. by @SarahScargall in #1214
• Add NebulaGraph Enterprise 5.2.1 Support for Episodic Memory by @wenhaocs in #1054
• Fix ty check type errors and ruff format violation by @Copilot in #1220

New Contributors

• @wenhaocs made their first contribution in #1054

Full Changelog: v0.3.0...v0.3.1

What's Changed

• Semantic Memory SetID refactor: customization by @o-love in #882
• Support basic CJK text sentence chunking by @edwinyyyu in #936
• add envvar MEMMACHINE_CONFIG_API by @jgong in #1020
• Support more search params for TypeScript client by @yaleMemVerge in #994
• Resolve ruff issues by @edwinyyyu in #1043
• Improve tools/chatgpt2memmachine migration tool to allow resume from previous run by @jgong in #1036
• [GH-1024] Fix set_id sql migration script by @o-love in #1027
• Define vector store interface by @edwinyyyu in #969
• Fix local failing tests and use a single flag for integration tests by @edwinyyyu in #1038
• tools/chatgpt2memmachine supports memory types by @jgong in #1053
• [CI] Add ty coverage to tests by @o-love in #1021
• Fix StrEnum usage for Python 3.10 compatibility by @edwinyyyu in #1056
• Add semantic memory management and episodic memory configuration API by @jealous in #1047
• Prefer docker compose than docker-compose in memmachine-compose.sh by @honggyukim in #1016
• update to n8n docs by @ChristianKniep in #1055
• Update README by @sscargal in #1044
• Bump json-repair from 0.55.0 to 0.55.2 by @dependabot[bot] in #1033
• Bump greenlet from 3.2.4 to 3.3.1 by @dependabot[bot] in #1032
• Bump prometheus-client from 0.23.1 to 0.24.1 by @dependabot[bot] in #1031
• Bump psutil from 7.1.3 to 7.2.2 by @dependabot[bot] in #1029
• Bump pytest from 9.0.1 to 9.0.2 by @dependabot[bot] in #1030
• Document more params for n8n integration by @yaleMemVerge in #1035
• set COMPOSE_CMD when running with commands by @jgong in #1067
• Update search result type definitions by @yaleMemVerge in #1059
• Optimize ChatGPT migration: disable STM during import, fix user_only filtering, and make Docker builds work on macOS. by @dxu104 in #1061
• Remove path manipulation antipatterns by @edwinyyyu in #1064
• Disable multiple process pytest by @edwinyyyu in #1072
• Docs: Updating Configuration parameters for n8n integration by @SarahScargall in #1046
• dedup citation IDs before validations by @jgong in #1099
• Change Local Docker Image message from WARNING to INFO by @sscargal in #1107
• Suppress Neo4j constraint notification spam in logs by @Copilot in #1106
• Remove Mypy by @edwinyyyu in #1084
• Bump boto3 from 1.42.21 to 1.42.44 by @dependabot[bot] in #1080
• Bump axios from 1.13.2 to 1.13.5 in /src/memmachine-ts/rest_client in the npm_and_yarn group across 1 directory by @dependabot[bot] in #1076
• Docs: Typescript updates to reflect new Parameters and Metrics method by @SarahScargall in #1058
• Bump langsmith from 0.4.53 to 0.6.3 in the uv group across 1 directory by @dependabot[bot] in #1075
• Reduce set_id pseudo-batch size to 5 by @o-love in #1085
• Limit should be min instead of max by @edwinyyyu in #1123
• Bump sqlalchemy from 2.0.44 to 2.0.46 by @dependabot[bot] in #1120
• Bump cohere from 5.20.1 to 5.20.5 by @dependabot[bot] in #1119
• Docs: New ChatGPT2MemMachine Tool Guide by @SarahScargall in #1114
• Bump fastmcp from 2.14.4 to 2.14.5 by @dependabot[bot] in #1078
• Bump json-repair from 0.55.2 to 0.57.1 by @dependabot[bot] in #1077
• Upgrade OpenAI dependency to v2.21.0 by @edwinyyyu in #1125
• Bump the uv group across 1 directory with 2 updates by @dependabot[bot] in #1127
• Switch uvicorn from websockets (deprecated) to websockets-sansio to resolve the deprecation warnings at startup by @sscargal in #1109
• fix(tests): replace wall-clock timing assertions with LLM call count … by @jealous in #1122
• Unify filtering behavior by @edwinyyyu in #1090
• Rest API: Search semantic memory metadata sets by @o-love in #1121
• Remove implicit prompt defaults from config by @o-love in #1138
• support helm chart deployment by @jgong in #1137
• Docs: Updating for new Config Endpoints, fixing typescript broken link by @SarahScargall in #1011
• Bump boto3 from 1.42.44 to 1.42.54 by @dependabot[bot] in #1134
• Docs: Updating PythonSDK Server content for new API Endpoints (Semantic Memory) by @SarahScargall in #1083
• Docs: Updates to Python SDK Client and REST API for new Semantic Memory Endpoints by @SarahScargall in #1074
• Bump pgvector from 0.4.1 to 0.4.2 by @dependabot[bot] in #1131
• Bump nltk from 3.9.2 to 3.9.3 in the uv group across 1 directory by @dependabot[bot] in #1150
• Bump fastmcp to 3.0.2 by @jealous in #1152
• [GH-1102] Clean invalid episode_ids during ingestion by @o-love in #1148
• Improve Profile Memory prompt by @o-love in #1082
• Bump ty from 0.0.14 to 0.0.20 by @jealous in #1153
• Bump cohere from 5.20.5 to 5.20.7 by @dependabot[bot] in #1157
• Bump json-repair from 0.57.1 to 0.58.2 by @dependabot[bot] in #1156
• Bump sentence-transformers from 5.2.0 to 5.2.3 by @dependabot[bot] in #1155
• [GH-1102] Fix dangling episode_ids by @o-love in #1151
• Add openclaw plugin by @malatewang in #1096
• Refactor packaging by @edwinyyyu in #1052
• Fix test data location by @edwinyyyu in #1167
• Retrieval agent implementation and benchmark updates by @Tianyang-Zhang in #1103
• [GH-1145] Reduce semantic memory log noise: demote routine trace logs to DEBUG by @o-love in #1162
• reduce docs docker image size by multi-stage build by @jgong in #1171
• Automate docs/openapi.json for the Python Server by @sscargal in #1168
• Fix broken paths by @edwinyyyu in #1173
• Bump the npm_and_yarn group across 1 directory with 2 updates by @dependabot[bot] in #1166
• Docs: Creation of new LlamaIndex Integration Guide by @SarahScargall in #1098
• Docs: Updating n8n Integration Guide to add Video Walkthrough by @SarahScargall in #1116
• Docs: New LangChain Integration Guide by @SarahScargall in #1097
• Add diagnostics for server by @edwinyyyu in #1135
• Docs: Updating Claude Code for Configuration Discrepancy by @SarahScargall in #1045
• Update definitions to support VectorStore implementations by @edwinyyyu in #1177
• [GH-1174] Fix episode_type enum UniqueViolationError during concurrent startup by @o-love in #1175
• Improve metrics by @edwinyyyu in #1176

New Contributors

• @Tianyang-Zhang made their first contribution in #1103

Full Changelog: v0.2.6...v0.3.0