Hindsight Files

What's Changed

• docs: changelog and blog post for v0.8.5 by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/2879
• Fix within-batch cosine similarity for non-unit embeddings by @koriyoshi2041 in https://github.com/vectorize-io/hindsight/pull/2890
• chore(deps): bump pillow, gitpython, pyasn1 (clears 62 high-severity alerts) by @dcbouius in https://github.com/vectorize-io/hindsight/pull/2899
• blog: Your 1M-Token Context Window Is Not Memory by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/2910
• feat(engine): filter list_memory_units by ingest age (created_before) by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/2902
• chore: regen skills/hindsight-docs openapi.json (unblocks verify-generated-files) by @dcbouius in https://github.com/vectorize-io/hindsight/pull/2925
• feat(ts-client): support custom headers by @menelsystemsbot in https://github.com/vectorize-io/hindsight/pull/2914
• chore(deps): bump npm transitive highs (brace-expansion, js-yaml, sharp, fast-uri, svgo, shell-quote) by @dcbouius in https://github.com/vectorize-io/hindsight/pull/2907
• blog: Give Roo Code a Memory So Every Task Builds on the Last by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/2931
• fix(engine): make mental-model refresh cutoff stubbable (fix red test-api on main) by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/2924
• fix backup restore schema compatibility by @oldnicke in https://github.com/vectorize-io/hindsight/pull/2920
• fix(retain): offset causal targets from the extraction-group start by @r266-tech in https://github.com/vectorize-io/hindsight/pull/2935
• chore(deps): bump protobuf to 7.x and OpenTelemetry to 1.44/0.65b0 by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/2923
• fix(llm): normalize bare LM Studio / Ollama base URL to /v1 (#2922) by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/2941
• feat(observability): diagnose blocked-loop vs pool-exhaustion on stalled /health by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/2942
• chore(deps): bump next, postcss, pypdf (clears the new high-severity wave) by @dcbouius in https://github.com/vectorize-io/hindsight/pull/2933
• fix(retain): stop append replay when an oversized doc's first transport slice extracts zero facts by @Korayem in https://github.com/vectorize-io/hindsight/pull/2930
• fix: propagate Codex reasoning effort by @Tosko4 in https://github.com/vectorize-io/hindsight/pull/2919
• fix(cli): preserve HTTP 400 details by @koriyoshi2041 in https://github.com/vectorize-io/hindsight/pull/2916
• feat(config): make store_document_text overridable per bank by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/2940
• feat(extensions): declare + provision extension-owned bank-scoped tables by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/2903
• fix(retain): make async retries idempotent via caller-supplied operation_id (#2937) by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/2947
• fix(repair): stop the shared-DB deadlock flake in test-api (CONCURRENTLY test DDL + retry transient deadlocks) by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/2943
• fix(mental-model): anchor delta refresh watermark to newest processed memory by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/2878
• feat(api): filter memory list by linked entity + entity timeline UI by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/2945
• fix(helm): override HINDSIGHT_API_PORT in worker StatefulSet to survive K8s service discovery by @Fyko in https://github.com/vectorize-io/hindsight/pull/2904
• feat(engine): add MemoryEngine.delete_memory_units bulk primitive by @chrislatimer in https://github.com/vectorize-io/hindsight/pull/2659
• feat(control-plane): display API version in sidebar footer by @handnewb in https://github.com/vectorize-io/hindsight/pull/2886
• fix(integrations): derive recall hook timeout from requestTimeoutSeconds by @handnewb in https://github.com/vectorize-io/hindsight/pull/2883
• fix(deps): require litellm>=1.93.0 for Python 3.14 support by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/2950
• feat(config): make embedding thresholds configurable by @Sanderhoff-alt in https://github.com/vectorize-io/hindsight/pull/2875
• docs: add TealTiger community integration listing by @nagasatish007 in https://github.com/vectorize-io/hindsight/pull/2831
• fix(auth): authorize bank writes before provisioning by @Sanderhoff-alt in https://github.com/vectorize-io/hindsight/pull/2646
• fix(curation): keep causal links across edit and invalidate/restore by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/2951
• docs(consolidation): define every input field in the consolidation prompt by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/2952
• feat(paperclip): per-agent enable/disable for pilot rollouts by @parties in https://github.com/vectorize-io/hindsight/pull/2724
• fix(oracle): unblock Oracle CI — free runner disk space + fix the retain deadlock by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/2948
• fix(graph): queue edited and restored memories for relinking by @Sanderhoff-alt in https://github.com/vectorize-io/hindsight/pull/2893
• blog: recall vs reflect (the two ways to read agent memory) by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/2954
• fix(files): reject misplaced retain metadata instead of discarding it by @r266-tech in https://github.com/vectorize-io/hindsight/pull/2971
• fix: accept text alias in fact extraction for schema-drift resilience by @jtstothard in https://github.com/vectorize-io/hindsight/pull/2969
• test(litellm): cover Responses named tool choice (#2953) by @ijevin in https://github.com/vectorize-io/hindsight/pull/2957
• fix: return free-form entities from dry-run extraction by @oldnicke in https://github.com/vectorize-io/hindsight/pull/2958
• docs(openclaw): note memory-wiki bridge mode is unsupported (#963) by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/2955
• fix(control-plane): preserve inherited observations configuration by @r266-tech in https://github.com/vectorize-io/hindsight/pull/2885
• fix(consolidation): stop emitting unsupported maxItems that breaks all Bedrock consolidation (#2500) by @r266-tech in https://github.com/vectorize-io/hindsight/pull/2502
• Add Russian temporal period rules by @dimonnld in https://github.com/vectorize-io/hindsight/pull/2767
• perf: eliminate redundant graph seed and UUID scans by @zommiommy in https://github.com/vectorize-io/hindsight/pull/2968
• fix(clients): expose async retain operation_id by @r266-tech in https://github.com/vectorize-io/hindsight/pull/2978
• fix: avoid dotenv side effects on library import by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/2979
• fix(control-plane): gate audit-log & observations tabs on resolved per-bank config by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/2982
• fix(oracle): don't COALESCE a bind against the CLOB mission column in update_bank by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/2981
• docs(zapier): clarify private-beta availability + Webhooks-by-Zapier path by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/2989
• blog: What people actually build with agent memory (use cases) by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/2990
• feat(zapier): remove memoryDefenseTriggered trigger (gated capability) by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/2994
• fix(worker): reconcile stranded batch_retain parents on recovery (#2985) by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/2986
• fix(claude-code): reflect tool calls + honor configured model (#2966, [#2881]) by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/2980
• fix(oracle): skip LLM trace writes on Oracle (llm_requests is PG-only) by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/3012
• fix(embeddings,reranker): default local models to CPU on Apple Silicon (MPS memory leak) by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/2988
• fix(oracle): audit_log write qualification + llm_requests read gating by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/3015
• fix(reflect): fail on unusable tool calls instead of salvaging leaked text by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/3013
• fix(transfer): preserve consolidation lifecycle on whole-bank import (#2965) by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/3017
• fix(operations): re-runnable batch_retain parents (retry re-queues children) by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/3018
• refactor(retain): require explicit semantic link thresholds by @Sanderhoff-alt in https://github.com/vectorize-io/hindsight/pull/3004
• fix(llm): send reasoning_effort on the tool path, matching call() by @merlinr68 in https://github.com/vectorize-io/hindsight/pull/2983
• fix(worker): stop wedged retains from holding worker slots forever by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/3020
• feat(copilot-cli): add GitHub Copilot CLI hooks integration by @ScottGuymer in https://github.com/vectorize-io/hindsight/pull/2742
• blog: How to move your agent's memory off a vector database by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/3022
• fix(tei): retry HTTP 429 backpressure for reranking and embeddings by @r266-tech in https://github.com/vectorize-io/hindsight/pull/3001
• docs+config(worker): rename per-type WORKER__MAX_SLOTS to _RESERVED_SLOTS (#2963) by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/3016
• fix(tracing): serialize unvalidated provider responses by @oldnicke in https://github.com/vectorize-io/hindsight/pull/3033
• chore(repo): remove unused code by @Sanderhoff-alt in https://github.com/vectorize-io/hindsight/pull/3007
• fix(oracle): fix three PG-isms breaking the Oracle backend by @dcbouius in https://github.com/vectorize-io/hindsight/pull/3021
• feat(config): per-operation reasoning_effort override (#2998) by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/3043
• fix(graph): serialize graph-maintenance queue enqueue against worker drain (#3034) by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/3045
• feat(retain): report zero-fact documents at write time (#3040) by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/3044
• fix(clients): expose mental model query controls by @r266-tech in https://github.com/vectorize-io/hindsight/pull/3042
• feat(reflect): add apply_all_directives to bypass directive tag scoping (#3031) by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/3046
• fix(reflect): apply exact empty scope to mental models by @Sanderhoff-alt in https://github.com/vectorize-io/hindsight/pull/3039
• fix(clients): expose mental model query controls in python wrapper by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/3047
• fix(control-plane): bind the dark: variant to the .dark class by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/3050
• feat(control-plane): add document tag filtering and unify facet chips by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/3049

New Contributors

• @menelsystemsbot made their first contribution in https://github.com/vectorize-io/hindsight/pull/2914
• @nagasatish007 made their first contribution in https://github.com/vectorize-io/hindsight/pull/2831
• @parties made their first contribution in https://github.com/vectorize-io/hindsight/pull/2724
• @jtstothard made their first contribution in https://github.com/vectorize-io/hindsight/pull/2969
• @merlinr68 made their first contribution in https://github.com/vectorize-io/hindsight/pull/2983
• @ScottGuymer made their first contribution in https://github.com/vectorize-io/hindsight/pull/2742

Full Changelog: https://github.com/vectorize-io/hindsight/compare/v0.8.5...v0.8.6