Skip to content
CAI
Software that uses CAICheck a score

xerj-org/xerj

41.6

Weak · 17 November 2026

564.2k

lines of production code

Rust

with Python, JavaScript

1

measurement over time

CAI band scale
CAI lens gauges

What this system is

Behavioural changes

1 commit (0 fixes) modifying benchmarks/agent-intents-2026-10-19, scripts/seo/intent\_covers

A change to existing behaviour in benchmarks/agent-intents-2026-10-19, scripts/seo/intent\_covers — 1 commit, 27 files: “answers: five agent-intent pages with light covers (\#1964)” (refs \#1964).

_benchmarks/agent-intents-2026-10-19, scripts/seo/intent\covers · medium confidence · unverified

1 commit (0 fixes) modifying benchmarks/agent-memory

A change to existing behaviour in benchmarks/agent-memory — 1 commit, 36 files: “seo: restore AI agent memory benchmark pages from \#2788 onto main (\#2946)” (refs \#2788, \#2946).

benchmarks/agent-memory · medium confidence · unverified

1 commit (0 fixes) modifying benchmarks/agent-queries-2026-10-20

A change to existing behaviour in benchmarks/agent-queries-2026-10-20 — 1 commit, 74 files: “GEO: eight answer pages from the 2026-10-20 agent-query captures (\#2251)” (refs \#2251).

benchmarks/agent-queries-2026-10-20 · medium confidence · unverified

1 commit (0 fixes) modifying benchmarks/beir-hybrid/results/2026-09-20-rerank-full

A change to existing behaviour in benchmarks/beir-hybrid/results/2026-09-20-rerank-full — 1 commit, 4 files: “bench: full 300-query Jev run — ranking up, calibration not there”.

benchmarks/beir-hybrid/results/2026-09-20-rerank-full · medium confidence · unverified

1 commit (0 fixes) modifying benchmarks/beir-hybrid/results/2026-09-20-rerank-pilot

A change to existing behaviour in benchmarks/beir-hybrid/results/2026-09-20-rerank-pilot — 1 commit, 4 files: “rerank: put each candidate inside its own noul — 0.3822 -\> 0.8299 nDCG@10”.

benchmarks/beir-hybrid/results/2026-09-20-rerank-pilot · medium confidence · unverified

1 commit (0 fixes) modifying benchmarks/beir-hybrid/results/2026-09-30-stemming-1059

A change to existing behaviour in benchmarks/beir-hybrid/results/2026-09-30-stemming-1059 — 1 commit, 10 files: “measure(beir-hybrid): stemming gate \#1059 — zero-hit 25-\>15 (FAIL), BM25 +0.0160/+0.0179 (PASS)” (refs \#1059).

benchmarks/beir-hybrid/results/2026-09-30-stemming-1059 · medium confidence · unverified

1 commit (0 fixes) modifying benchmarks/best-of-jsonl-search-2026, scripts/seo/bestof\_covers

A change to existing behaviour in benchmarks/best-of-jsonl-search-2026, scripts/seo/bestof\_covers — 1 commit, 11 files: “compare: four best-of-2026 pages with light covers (\#1968)” (refs \#1968).

_benchmarks/best-of-jsonl-search-2026, scripts/seo/bestof\covers · medium confidence · unverified

1 commit (0 fixes) modifying benchmarks/decide-model

A change to existing behaviour in benchmarks/decide-model — 1 commit, 19 files: “decide: open xerj-decide model — training harness, artifact, eval card (\#1064)” (refs \#1064).

benchmarks/decide-model · medium confidence · unverified

1 commit (0 fixes) modifying benchmarks/grep-vs-retrieval-lucene

A change to existing behaviour in benchmarks/grep-vs-retrieval-lucene — 1 commit, 15 files: “benchmarks: grep-vs-retrieval harness for Lucene (no results yet)”.

benchmarks/grep-vs-retrieval-lucene · medium confidence · unverified

1 commit (0 fixes) modifying benchmarks/ingest-neural-bulk

A change to existing behaviour in benchmarks/ingest-neural-bulk — 1 commit, 4 files: “feat(bench): neural-bulk A/B harness and frozen Okta corpus for ingest changes (\#2448)” (refs \#2448).

benchmarks/ingest-neural-bulk · medium confidence · unverified

1 commit (0 fixes) modifying benchmarks/map-gate

A change to existing behaviour in benchmarks/map-gate — 1 commit, 69 files: “benchmarks/map-gate: measure the xerj\_map unknown-field-400 gate (\#1055) — 20/30 -\> 0/30, PASS” (refs \#1055).

benchmarks/map-gate · medium confidence · unverified

1 commit (0 fixes) modifying benchmarks/skills-intent

A change to existing behaviour in benchmarks/skills-intent — 1 commit, 5 files: “feat(autoindex): --skills finds and indexes skill folders (\#2644) (\#2802)” (refs \#2644, \#2802).

benchmarks/skills-intent · medium confidence · unverified

1 commit (0 fixes) modifying benchmarks/systemone-classify-email

A change to existing behaviour in benchmarks/systemone-classify-email — 1 commit, 5 files: “bench(systemone): measure local /\_decide for discussion-1012 email labelling”.

benchmarks/systemone-classify-email · medium confidence · unverified

1 commit (0 fixes) modifying benchmarks/systemone-watcher-gate

A change to existing behaviour in benchmarks/systemone-watcher-gate — 1 commit, 7 files: “detections(\#1108): run the \#1062 wrong-and-confident gate through the real \_watcher path — it FAILS, published” (refs \#1108, \#1062).

benchmarks/systemone-watcher-gate · medium confidence · unverified

1 commit (0 fixes) modifying demo/data

A change to existing behaviour in demo/data — 1 commit, 5 files: “docs(recipes): all-way search + autoindex→neural semantic search”.

demo/data · medium confidence · unverified

1 commit (0 fixes) modifying demo/playbooks/neural-embedder-verification

A change to existing behaviour in demo/playbooks/neural-embedder-verification — 1 commit, 8 files: “chore(playbooks): reproducible neural-embedder verification harness”.

demo/playbooks/neural-embedder-verification · medium confidence · unverified

1 commit (0 fixes) modifying demo/playbooks/onnx-model-optimization

A change to existing behaviour in demo/playbooks/onnx-model-optimization — 1 commit, 6 files: “perf(onnx): add reproducible fused FP32 model recipe (\#154)” (refs \#154).

demo/playbooks/onnx-model-optimization · medium confidence · unverified

1 commit (0 fixes) modifying demo/usecases/autoindex/recipe-captures

A change to existing behaviour in demo/usecases/autoindex/recipe-captures — 1 commit, 17 files: “docs+site: AI-first value prop — use cases lead, autoindex flagship”.

demo/usecases/autoindex/recipe-captures · medium confidence · unverified

1 commit (0 fixes) modifying demo/usecases/autoindex/sim-transcripts

A change to existing behaviour in demo/usecases/autoindex/sim-transcripts — 1 commit, 45 files: “feat(autoindex): zero-config folder discovery + indexing as a built-in subcommand”.

demo/usecases/autoindex/sim-transcripts · medium confidence · unverified

1 commit (0 fixes) modifying engine/crates/xerj-api/examples

A change to existing behaviour in engine/crates/xerj-api/examples — 1 commit, 1 file: “perf(binary-protocol): \_bulk shares one embedding pass across the batch (\#903) (\#912)” (refs \#903, \#912).

engine/crates/xerj-api/examples · medium confidence · unverified

1 commit (0 fixes) modifying engine/crates/xerj-server

A change to existing behaviour in engine/crates/xerj-server — 1 commit, 1 file: “feat(grpc): implement real tonic XerjSearch server on :8081”.

engine/crates/xerj-server · medium confidence · unverified

1 commit (0 fixes) modifying integrations

A change to existing behaviour in integrations — 1 commit, 5 files: “feat: add Herdr integration (\#2891)” (refs \#2891).

integrations · medium confidence · unverified

1 commit (0 fixes) modifying playground

A change to existing behaviour in playground — 1 commit, 5 files: “Xerj monorepo: engine, Xerj Console, landing, demo, brief”.

playground · medium confidence · unverified

1 commit (0 fixes) modifying scripts/ci

A change to existing behaviour in scripts/ci — 1 commit, 2 files: “ci: skip the heavy Rust jobs on docs-only pull requests (fail open) (\#1477)” (refs \#1477).

scripts/ci · medium confidence · unverified

1 commit (0 fixes) modifying scripts/git-hooks

A change to existing behaviour in scripts/git-hooks — 1 commit, 4 files: “chore(git): identity-guard hooks — stop concurrent-session identity leaks at commit time”.

scripts/git-hooks · medium confidence · unverified

1 commit (0 fixes) modifying scripts/perf

A change to existing behaviour in scripts/perf — 1 commit, 5 files: “perf: release performance gate for autoindex + search (budgets, ES floor, rc.99 regression harness) (\#2432)” (refs \#2432).

scripts/perf · medium confidence · unverified

1 commit (0 fixes) modifying tools/agents

A change to existing behaviour in tools/agents — 1 commit, 1 file: “docs(examples): Cohere Embed, Command RAG with file citations, Cohere tool schema”.

tools/agents · medium confidence · unverified

1 commit (0 fixes) modifying tools/map-report, tools/map-report/fixture

A change to existing behaviour in tools/map-report, tools/map-report/fixture — 1 commit, 9 files: “feat(autoindex): export offline HTML and Markdown data maps (\#2705)” (refs \#2705).

tools/map-report, tools/map-report/fixture · medium confidence · unverified

1 commit (0 fixes) modifying tools/session-memory

A change to existing behaviour in tools/session-memory — 1 commit, 3 files: “feat(cli): add opt-in Claude Code session handoff hooks (\#2849)” (refs \#2849).

tools/session-memory · medium confidence · unverified

10 commits (3 fixes) modifying engine/crates/xerj-rerank/src

A change to existing behaviour in engine/crates/xerj-rerank/src — 10 commits (3 fixs), 2 files: “ci: pass the fmt and clippy gates — first run was on the pushed branch”; “feat(api): POST /v1/systemone + /\_decide — the local judge speaks the System One wire”; “search: opt-in local \judge\ stage on \_search — \_p\_relevant per hit (\#1060)”; and 7 more (refs \#1060, \#968, \#946, \#2635, \#2889).

engine/crates/xerj-rerank/src · medium confidence · unverified

11 commits (2 fixes) modifying recipes

A change to existing behaviour in recipes — 11 commits (2 fixs), 11 files: “fix(engine,vector): \#371 scalar8 kNN scored updated docs from stale codes AND a stale codebook”; “docs(recipes): all-way search + autoindex→neural semantic search”; “knn: scalar8 codes written at ingest, slot-addressed (\#392)”; and 8 more (refs \#371, \#392, \#2377).

recipes · medium confidence · unverified

11 commits (3 fixes) modifying landing/playground

A change to existing behaviour in landing/playground — 11 commits (3 fixs), 5 files: “fix(playground): restore saved views without stale URL state (\#2094, from \#2071)”; “fix(playground): resolve custom dashboard templates consistently (\#2088, from \#2072)”; “docs(knn): kNN claims corrected in both directions — HNSW-served ANN with exact rescoring, measured recall, exact fallback”; and 8 more (refs \#2094, \#2071, \#2088, \#2072, \#2323, …).

landing/playground · medium confidence · unverified

12 commits (1 fix) modifying demo/playbooks/rc12

A change to existing behaviour in demo/playbooks/rc12 — 12 commits (1 fix), 13 files: “test(perf): record the doc-values denominator — the lever is a range, not 37%”; “test(perf): ES-YAML conformance passes with both rc.12 engine changes”; “polish(scripts): rc12 harness comment no longer says --port does not exist (\#1558)”; and 9 more (refs \#1558, \#1554).

demo/playbooks/rc12 · medium confidence · unverified

12 commits (4 fixes) modifying engine/crates/xerj-storage/src/stored\_codec

A change to existing behaviour in engine/crates/xerj-storage/src/stored\_codec — 12 commits (4 fixs), 2 files: “Cut cold first-query search latency (\#2422) (\#2443)”; “storage: don't persist derived semantic\_text vectors in stored \_source (\#1122, \#2433) (\#2471)”; “fix(storage): preserve explicit \_source nulls through the columnar codec (\#415)”; and 9 more (refs \#2422, \#2443, \#1122, \#2433, \#2471, …).

_engine/crates/xerj-storage/src/stored\codec · medium confidence · unverified

13 commits (4 fixes) modifying engine/crates/xerj-console-api/src/auth

A change to existing behaviour in engine/crates/xerj-console-api/src/auth — 13 commits (4 fixs), 10 files: “fix(console-api): require a passkey session to mint API tokens ([CVE redacted]) (\#2003)”; “fix(console-api): fail closed on view delete and auth upserts”; “security: Phase-2 hardening — snapshot escape, field-limit, magic-link TOCTOU, info disclosure, index-name boundary; perf: ONNX windowing”; and 10 more (refs \#2003, \#76, \#112, \#935).

engine/crates/xerj-console-api/src/auth · medium confidence · unverified

14 commits (7 fixes) modifying tools/hub-packs

A change to existing behaviour in tools/hub-packs — 14 commits (7 fixs), 6 files: “ci(hub): the whole-hub end-to-end gate as a runnable script + workflow (spec §12.4) (\#1725)”; “fix(hub-packs): isolate malformed manifest shapes (\#2099, from \#2065)”; “fix(hub-e2e): make the sandboxed gate run on a CI runner (\#1876)”; and 11 more (refs \#1725, \#2099, \#2065, \#1876, \#1582, …).

tools/hub-packs · medium confidence · unverified

16 commits (2 fixes) modifying content/compare

A change to existing behaviour in content/compare — 16 commits (2 fixs), 29 files: “Add six category pages: hybrid search, BM25 vs vector, Mem0, claude-mem and MemPalace, Claude Code memory, agentic RAG (\#2785)”; “fix(seo): require updated: frontmatter in every article source (\#974)”; “site: canonical description, synced figures, FAQPage and MCP listing manifests (\#2323)”; and 13 more (refs \#2785, \#974, \#2323, \#2788, \#2946, …).

content/compare · medium confidence · unverified

16 commits (5 fixes) modifying python

A change to existing behaviour in python — 16 commits (5 fixs), 3 files: “feat(share): share links — CLI, guest reading room, tunnel (\#947)”; “ci(hub): the whole-hub end-to-end gate as a runnable script + workflow (spec §12.4) (\#1725)”; “fix(hub-packs): isolate malformed manifest shapes (\#2099, from \#2065)”; and 13 more (refs \#947, \#1725, \#2099, \#2065, \#974, …).

python · medium confidence · unverified

17 commits (2 fixes) modifying landing/demo

A change to existing behaviour in landing/demo — 17 commits (2 fixs), 1 file: “docs(landing): verify release downloads on upgrades and demo pages (\#454)”; “release: cut v1.0.0-rc.5”; “site: canonical description, synced figures, FAQPage and MCP listing manifests (\#2323)”; and 14 more (refs \#454, \#2323, \#492, \#1615, \#1920, …).

landing/demo · medium confidence · unverified

18 commits (6 fixes) modifying engine/crates/xerj-cluster/src

A change to existing behaviour in engine/crates/xerj-cluster/src — 18 commits (6 fixs), 15 files: “feat(cluster): replicate acknowledged document writes onto shard replicas (\#2287)”; “feat(cluster): route \_bulk and \_update writes to the shard primary (\#2344)”; “fix(engine): degrade when a shard thread or merge buffer cannot be allocated (\#1122) (\#2315)”; and 15 more (refs \#2287, \#2344, \#1122, \#2315, \#2257, …).

engine/crates/xerj-cluster/src · medium confidence · unverified

2 commits (0 fixes) modifying benchmarks/ask-plan

A change to existing behaviour in benchmarks/ask-plan — 2 commits, 25 files: “bench(ask-plan): latency arm — the \#1056 p50 gate, measured 1.129 ms”; “feat(bench): ask-plan harness skeleton for the \#1056 gate — data + scripts, zero measured results” (refs \#1056).

benchmarks/ask-plan · medium confidence · unverified

2 commits (0 fixes) modifying benchmarks/beir-hybrid/results/2026-09-30-judge-gate

A change to existing behaviour in benchmarks/beir-hybrid/results/2026-09-30-judge-gate — 2 commits, 12 files: “bench: raw log for the FiQA bm30+judge arm — re-run clean, 0.1650 x3 (spread 0.0000)”; “beir-hybrid: local-judge gate \#1060 measured — FAIL on quality, PASS on cost” (refs \#1060).

benchmarks/beir-hybrid/results/2026-09-30-judge-gate · medium confidence · unverified

2 commits (0 fixes) modifying brief, brief/kit

A change to existing behaviour in brief, brief/kit — 2 commits, 69 files: “docs: OSS-release scaffolding (README, LICENSE, CONTRIBUTING, CoC, SECURITY, templates)”; “Xerj monorepo: engine, Xerj Console, landing, demo, brief”.

brief, brief/kit · medium confidence · unverified

2 commits (0 fixes) modifying demo/agent-gate

A change to existing behaviour in demo/agent-gate — 2 commits, 8 files: “docs(demo): real AI-agent usage scenarios; deprecate stale sales kit”; “test(agent-gate): honest measurement of whether XERJ helps an AI agent”.

demo/agent-gate · medium confidence · unverified

2 commits (0 fixes) modifying demo/scripts

A change to existing behaviour in demo/scripts — 2 commits, 1 file: “polish(scripts): demo setup.sh passes XERJ\_ES\_PORT to the server it polls (\#1548)”; “Xerj monorepo: engine, Xerj Console, landing, demo, brief” (refs \#1548).

demo/scripts · medium confidence · unverified

2 commits (0 fixes) modifying demo/usecases/doc-index

A change to existing behaviour in demo/usecases/doc-index — 2 commits, 9 files: “usecase(doc-index): Claude+XERJ vs grep-only agent on a recursive doc folder”; “perf(read): eliminate O(N) read cliffs + columnar stats aggs (ES-YAML 1360/0)”.

demo/usecases/doc-index · medium confidence · unverified

2 commits (0 fixes) modifying engine/crates/xerj-engine/benches

A change to existing behaviour in engine/crates/xerj-engine/benches — 2 commits, 1 file: “chore(lint): clear all 146 outstanding clippy + dead-code lints”; “Xerj monorepo: engine, Xerj Console, landing, demo, brief”.

engine/crates/xerj-engine/benches · medium confidence · unverified

2 commits (0 fixes) modifying engine/crates/xerj-engine/examples

A change to existing behaviour in engine/crates/xerj-engine/examples — 2 commits, 1 file: “chore(lint): clear all 146 outstanding clippy + dead-code lints”; “perf(memtable): intern doc-id as Arc\<str\> — kill the per-token String malloc under the shard write lock”.

engine/crates/xerj-engine/examples · medium confidence · unverified

2 commits (0 fixes) modifying tools/packs/keys

A change to existing behaviour in tools/packs/keys — 2 commits, 3 files: “pack: agent-session-trajectories on main (mirror of corpus-hub fc8f86d82 + 1cfad62d1)”; “feat(autoindex): signed corpus packs — keygen/sign/--verify-sig + scheduled rust-vulns publish (M6)”.

tools/packs/keys · medium confidence · unverified

2 commits (0 fixes) modifying tools/search-hints

A change to existing behaviour in tools/search-hints — 2 commits, 4 files: “feat(cli): add opt-in Claude Code search hints (\#2869)”; “feat(cli): add opt-in Gemini CLI search hints (\#2879)” (refs \#2869, \#2879).

tools/search-hints · medium confidence · unverified

2 commits (1 fix) modifying benchmarks/decisions-calibration

A change to existing behaviour in benchmarks/decisions-calibration — 2 commits (1 fix), 11 files: “systemone: calibration layer — p\_cal beside every p\_raw (\#1063)”; “decisions-calibration: \#1063 rc.80 final form — pair-level FiQA gate PASS 0.0088; PAVA mirror fix” (refs \#1063).

benchmarks/decisions-calibration · medium confidence · unverified

2 commits (1 fix) modifying benchmarks/hybrid-scan-939

A change to existing behaviour in benchmarks/hybrid-scan-939 — 2 commits (1 fix), 7 files: “fix(knn): clippy 1.97 clean for the \#939 scan; latency harness records what it timed”; “perf(knn): exact scan ranks addresses and hydrates only the winners (\#939)” (refs \#939).

benchmarks/hybrid-scan-939 · medium confidence · unverified

2 commits (1 fix) modifying benchmarks/neural-path-triage

A change to existing behaviour in benchmarks/neural-path-triage — 2 commits (1 fix), 23 files: “fix(ingest): bounded window concurrency for neural \_bulk embedding (\#938 slice 1)”; “docs(roadmap): the zero-token direction (\#942)” (refs \#938, \#942).

benchmarks/neural-path-triage · medium confidence · unverified

2 commits (1 fix) modifying benchmarks/systemone-gate

A change to existing behaviour in benchmarks/systemone-gate — 2 commits (1 fix), 5 files: “feat(api): POST /v1/systemone + /\_decide — the local judge speaks the System One wire”; “fix(api): /v1/systemone votes on the payload, never the instruction prose”.

benchmarks/systemone-gate · medium confidence · unverified

2 commits (1 fix) modifying engine/crates/xerj-sources

A change to existing behaviour in engine/crates/xerj-sources — 2 commits (1 fix), 13 files: “feat(sources): live Postgres sources: consistent snapshot, then logical replication”; “fix(cli): one \xerj sources\ command: live sources, \--types\ for the catalogue”.

engine/crates/xerj-sources · medium confidence · unverified

2 commits (1 fix) modifying engine/crates/xerj-storage/examples

A change to existing behaviour in engine/crates/xerj-storage/examples — 2 commits (1 fix), 2 files: “fix(storage): harden bounded selective decode”; “Cold FTS top-k: walk on identities, hydrate only the page's winners (\#2454)” (refs \#2454).

engine/crates/xerj-storage/examples · medium confidence · unverified

2 commits (1 fix) modifying engine/fuzz

A change to existing behaviour in engine/fuzz — 2 commits (1 fix), 82 files: “fix(security): finish the \#207 date-math fix and close the same defect in the agg script tokenisers”; “feat(security): make the cargo-audit and cargo-fuzz claims real, and fix what they found (\#207)” (refs \#207).

engine/fuzz · medium confidence · unverified

2 commits (1 fix) modifying tools/factory-definitions

A change to existing behaviour in tools/factory-definitions — 2 commits (1 fix), 2 files: “fix(autoindex): capture separately exported factory constants”; “test(autoindex): record factory definition lifecycle and gate evidence”.

tools/factory-definitions · medium confidence · unverified

23 commits (9 fixes) modifying engine/crates/xerj-vector

A change to existing behaviour in engine/crates/xerj-vector — 23 commits (9 fixs), 11 files: “style: clear the remaining clippy 1.97 lints (-D warnings green)”; “docs(vector): correct insert\_batch doc — it is serial, not rayon-parallel”; “fix(vector): compact the HNSW graph when overwrites leave it mostly dead (\#2799)”; and 20 more (refs \#2799, \#2809, \#371, \#2812, \#392, …).

engine/crates/xerj-vector · medium confidence · unverified

24 commits (0 fixes) modifying landing/architecture, landing/security

A change to existing behaviour in landing/architecture, landing/security — 24 commits, 2 files: “landing: keep the Day/Night choice across pages (\#2169)”; “site: Cohere integration page, on-prem worked example, docs/COHERE.md, COHERE nav item”; “site: answers in the header menu, hub-first footer, no brand links (\#2410)”; and 21 more (refs \#2169, \#2410, \#2323, \#1352, \#2311, …).

landing/architecture, landing/security · medium confidence · unverified

26 commits (8 fixes) modifying scripts

A change to existing behaviour in scripts — 26 commits (8 fixs), 14 files: “feat(feedback): xerj feedback command + narrow field-report CLA carve-out”; “chore(release): run the artifact on Apple Silicon, and fail closed when nothing ran”; “Complete the MCP compatibility and executable contract audit”; and 23 more (refs \#320, \#1808, \#949, \#2096, \#2073, …).

scripts · medium confidence · unverified

27 commits (10 fixes) modifying xerj-ux

A change to existing behaviour in xerj-ux — 27 commits (10 fixs), 95 files: “feat(share): share links — CLI, guest reading room, tunnel (\#947)”; “feat(console): corpus home, a reader for any record, read-only guest mode (supersedes \#923) (\#945)”; “feat(autoindex): sharedterm@1 — link documents by the distinctive words they share”; and 23 more (refs \#947, \#923, \#945, \#1425, \#210, …).

xerj-ux · medium confidence · unverified

27 commits (3 fixes) modifying landing/brandbook

A change to existing behaviour in landing/brandbook — 27 commits (3 fixs), 2 files: “landing: keep the Day/Night choice across pages (\#2169)”; “site: Cohere integration page, on-prem worked example, docs/COHERE.md, COHERE nav item”; “site: answers in the header menu, hub-first footer, no brand links (\#2410)”; and 24 more (refs \#2169, \#2410, \#2323, \#1352, \#2311, …).

landing/brandbook · medium confidence · unverified

29 commits (4 fixes) modifying landing/resources

A change to existing behaviour in landing/resources — 29 commits (4 fixs), 2 files: “landing: keep the Day/Night choice across pages (\#2169)”; “site: Cohere integration page, on-prem worked example, docs/COHERE.md, COHERE nav item”; “site: answers in the header menu, hub-first footer, no brand links (\#2410)”; and 26 more (refs \#2169, \#2410, \#2323, \#492, \#1352, …).

landing/resources · medium confidence · unverified

3 commits (0 fixes) modifying content/visuals

A change to existing behaviour in content/visuals — 3 commits, 36 files: “Add six category pages: hybrid search, BM25 vs vector, Mem0, claude-mem and MemPalace, Claude Code memory, agentic RAG (\#2785)”; “seo: restore AI agent memory benchmark pages from \#2788 onto main (\#2946)”; “Rebuild 5 low-traffic answers pages with diagrams, charts and a small interactive (\#2742)” (refs \#2785, \#2788, \#2946, \#2742).

content/visuals · medium confidence · unverified

3 commits (0 fixes) modifying demo/usecases/second-brain

A change to existing behaviour in demo/usecases/second-brain — 3 commits, 5 files: “ci(usecases): gate the second-brain, MCP and autoindex harnesses”; “feat(second-brain): verified brain-CLI slice — demo corpus, MCP brain tools, sniffer + harness fixes”; “polish(scripts): second-brain transcript.sh stops sending -- as a query pair (\#1550)” (refs \#1550).

demo/usecases/second-brain · medium confidence · unverified

3 commits (0 fixes) modifying engine/crates/xerj-logs

A change to existing behaviour in engine/crates/xerj-logs — 3 commits, 5 files: “chore(lint): clear Rust 1.96 clippy lints (CI toolchain parity)”; “chore(lint): clear all 146 outstanding clippy + dead-code lints”; “Xerj monorepo: engine, Xerj Console, landing, demo, brief”.

engine/crates/xerj-logs · medium confidence · unverified

3 commits (0 fixes) modifying engine/proto

A change to existing behaviour in engine/proto — 3 commits, 1 file: “feat(hub): the gRPC XerjHub service, a thin adapter over the /\_hub/v1 code (plan §7)”; “Xerj monorepo: engine, Xerj Console, landing, demo, brief”; “feat(hub-v2): lock the XerjHub gRPC contract in the proto”.

engine/proto · medium confidence · unverified

3 commits (1 fix) modifying benchmarks/hub-pov

A change to existing behaviour in benchmarks/hub-pov — 3 commits (1 fix), 15 files: “bench(hub-pov): runnable harness for the Hub v2 proof of value + smoke run (\#1722)”; “fix(hub-e2e): make the sandboxed gate run on a CI runner (\#1876)”; “docs(hub-pov): pre-register the Corpus Hub v2 proof-of-value measurement” (refs \#1722, \#1876).

benchmarks/hub-pov · medium confidence · unverified

3 commits (1 fix) modifying demo/usecases/autoindex/scale

A change to existing behaviour in demo/usecases/autoindex/scale — 3 commits (1 fix), 12 files: “feat(autoindex): zero-config folder discovery + indexing as a built-in subcommand”; “fix(engine): bound all segment hydration cache owners”; “test(ingest): add bounded shipped-path diagnostics”.

demo/usecases/autoindex/scale · medium confidence · unverified

30 commits (13 fixes) modifying engine/crates/xerj-ai

A change to existing behaviour in engine/crates/xerj-ai — 30 commits (13 fixs), 13 files: “style: clear the remaining clippy 1.97 lints (-D warnings green)”; “feat(ai): auto-embed-on-ingest for semantic\_text (zero-config built-in embedder)”; “storage: don't persist derived semantic\_text vectors in stored \_source (\#1122, \#2433) (\#2471)”; and 27 more (refs \#1122, \#2433, \#2471, \#2807, \#2860, …).

engine/crates/xerj-ai · medium confidence · unverified

31 commits (3 fixes) modifying landing/industries

A change to existing behaviour in landing/industries — 31 commits (3 fixs), 5 files: “landing: keep the Day/Night choice across pages (\#2169)”; “site: Cohere integration page, on-prem worked example, docs/COHERE.md, COHERE nav item”; “site: answers in the header menu, hub-first footer, no brand links (\#2410)”; and 28 more (refs \#2169, \#2410, \#2323, \#491, \#1352, …).

landing/industries · medium confidence · unverified

36 commits (10 fixes) modifying engine/crates/xerj-console-api/src

A change to existing behaviour in engine/crates/xerj-console-api/src — 36 commits (10 fixs), 17 files: “feat(share): share links — CLI, guest reading room, tunnel (\#947)”; “fix(console-api): require a passkey session to mint API tokens ([CVE redacted]) (\#2003)”; “feat(console): corpus home, a reader for any record, read-only guest mode (supersedes \#923) (\#945)”; and 33 more (refs \#947, \#2003, \#923, \#945, \#210, …).

engine/crates/xerj-console-api/src · medium confidence · unverified

38 commits (10 fixes) modifying landing/case-studies

A change to existing behaviour in landing/case-studies — 38 commits (10 fixs), 5 files: “landing: keep the Day/Night choice across pages (\#2169)”; “site: Cohere integration page, on-prem worked example, docs/COHERE.md, COHERE nav item”; “site: answers in the header menu, hub-first footer, no brand links (\#2410)”; and 35 more (refs \#2169, \#2410, \#2323, \#492, \#1352, …).

landing/case-studies · medium confidence · unverified

384 commits (62 fixes) modifying landing

A change to existing behaviour in landing — 384 commits (62 fixs), 39 files: “chore(seo): sitemap lastmod regenerated post-commit (own-commit rule)”; “landing: keep the Day/Night choice across pages (\#2169)”; “feat(share): share links — CLI, guest reading room, tunnel (\#947)”; and 372 more (refs \#2169, \#947, \#1122, \#923, \#945, …).

landing · medium confidence · unverified

4 commits (0 fixes) modifying benchmarks/index-size

A change to existing behaviour in benchmarks/index-size — 4 commits, 9 files: “fts: ZFM5 meta records — columnar varint streams under the zstd envelope”; “benchmarks/index-size: measure the force-merged on-disk footprint per component”; “storage: ZBS4 stored-section slim — typed ints, copy-of refs, merge-path zstd chooser”; and 1 more.

benchmarks/index-size · medium confidence · unverified

4 commits (0 fixes) modifying tools/packs/xerj-blogposts

A change to existing behaviour in tools/packs/xerj-blogposts — 4 commits, 12 files: “release(rc.91): version stamp, CHANGELOG, ROADMAP roll, landing stamps”; “Blog: the exploit hub, tested on three fresh CVEs (+ blogposts corpus to 8 records)”; “blog: one cover series for every post (field-log posters, og + hero + thumbnail) (\#1918)”; and 1 more (refs \#1918, \#1915).

tools/packs/xerj-blogposts · medium confidence · unverified

4 commits (1 fix) modifying benchmarks/mbox-ingest

A change to existing behaviour in benchmarks/mbox-ingest — 4 commits (1 fix), 24 files: “feat(autoindex): mbox and Google Takeout ingest (\#949)”; “fix(engine): \#948 — prune published memtable entries, charge what they retain, bound merge batches, drain on breaker”; “polish(scripts): synth\_bulk\_ingest --help says --json is a no-op (\#1534)”; and 1 more (refs \#949, \#948, \#1534, \#1533).

benchmarks/mbox-ingest · medium confidence · unverified

4 commits (1 fix) modifying tools/catalog-map

A change to existing behaviour in tools/catalog-map — 4 commits (1 fix), 3 files: “test(map): publish generated scope schema and auth controls”; “fix(map): explain autoindex and managed corpus namespace boundaries”; “Record independent catalog validation and gate hybrid pagination in ES-YAML”; and 1 more.

tools/catalog-map · medium confidence · unverified

4 commits (2 fixes) modifying benchmarks/corpus-ab-2026-10

A change to existing behaviour in benchmarks/corpus-ab-2026-10 — 4 commits (2 fixs), 67 files: “bench(corpus-ab): record corpus-x100 design + measured S1 sample validation”; “bench(corpus-ab): \#1111 A/B results — 33 confirmed defect classes, corpus tie published”; “bench(corpus-ab): freeze the \#1111 A/B protocol before any run”; and 1 more (refs \#1111).

benchmarks/corpus-ab-2026-10 · medium confidence · unverified

4 commits (2 fixes) modifying deploy

A change to existing behaviour in deploy — 4 commits (2 fixs), 7 files: “fix(server): bind loopback by default, refuse a cleartext network bind (\#228)”; “fix(helm): default to secure, and say what insecure mode turns off”; “Xerj monorepo: engine, Xerj Console, landing, demo, brief”; and 1 more (refs \#228, \#1922).

deploy · medium confidence · unverified

4 commits (2 fixes) modifying engine/crates/xerj-hub/src/sync

A change to existing behaviour in engine/crates/xerj-hub/src/sync — 4 commits (2 fixs), 1 file: “feat(hub): xerj hub sync and sync --install-timer (plan §8)”; “test(hub): pin the systemd calendar check to a fixed clock”; “hub: one XERJ\_HUB\_PUBKEY test lock, restore the deleted cli test; timer home uses USERPROFILE (\#1706)”; and 1 more (refs \#1706, \#2249).

engine/crates/xerj-hub/src/sync · medium confidence · unverified

4 commits (2 fixes) modifying functions/api

A change to existing behaviour in functions/api — 4 commits (2 fixs), 1 file: “fix(field-report): raise verdict cap from 80 to 1000 (\#1847)”; “feat(functions): add /api/field-report HTTP fallback for agent field reports (\#1760)”; “fix(field-report): enforce the raw request byte limit (\#2845)”; and 1 more (refs \#1847, \#1760, \#2845).

functions/api · medium confidence · unverified

4 commits (2 fixes) modifying landing/playground/src/data

A change to existing behaviour in landing/playground/src/data — 4 commits (2 fixs), 4 files: “fix(playground): resolve custom dashboard templates consistently (\#2088, from \#2072)”; “fix(playground): retain query marker paint in PNG exports (\#2092, from \#2074)”; “Xerj monorepo: engine, Xerj Console, landing, demo, brief”; and 1 more (refs \#2088, \#2072, \#2092, \#2074).

landing/playground/src/data · medium confidence · unverified

4 commits (2 fixes) modifying tools/xerj-code/swebench

A change to existing behaviour in tools/xerj-code/swebench — 4 commits (2 fixs), 4 files: “fix(bench): refuse failed SWE-bench clones and pinned checkouts (\#2098, from \#2062)”; “fix(bench): preserve complete agent patches against the original task base (\#2095, from \#2063)”; “feat(\#977): port the reference-coding loop into the binary — xerj code, xerj corpus, xerj\_code\_search MCP”; and 1 more (refs \#2098, \#2062, \#2095, \#2063, \#977).

tools/xerj-code/swebench · medium confidence · unverified

42 commits (6 fixes) modifying landing/use-cases

A change to existing behaviour in landing/use-cases — 42 commits (6 fixs), 10 files: “landing: keep the Day/Night choice across pages (\#2169)”; “site: Cohere integration page, on-prem worked example, docs/COHERE.md, COHERE nav item”; “site: answers in the header menu, hub-first footer, no brand links (\#2410)”; and 39 more (refs \#2169, \#2410, \#2323, \#1352, \#371, …).

landing/use-cases · medium confidence · unverified

46 commits (5 fixes) modifying landing/compare

A change to existing behaviour in landing/compare — 46 commits (5 fixs), 58 files: “landing: keep the Day/Night choice across pages (\#2169)”; “feat(share): share links — CLI, guest reading room, tunnel (\#947)”; “feat(console): corpus home, a reader for any record, read-only guest mode (supersedes \#923) (\#945)”; and 43 more (refs \#2169, \#947, \#923, \#945, \#2410, …).

landing/compare · medium confidence · unverified

46 commits (7 fixes) modifying landing/benchmarks

A change to existing behaviour in landing/benchmarks — 46 commits (7 fixs), 2 files: “landing: keep the Day/Night choice across pages (\#2169)”; “site: Cohere integration page, on-prem worked example, docs/COHERE.md, COHERE nav item”; “site: answers in the header menu, hub-first footer, no brand links (\#2410)”; and 43 more (refs \#2169, \#2410, \#2323, \#1352, \#1920, …).

landing/benchmarks · medium confidence · unverified

48 commits (24 fixes) modifying engine/crates/xerj-hub/src

A change to existing behaviour in engine/crates/xerj-hub/src — 48 commits (24 fixs), 15 files: “feat(hub): Corpus Hub v2 keystone — pack build/pull + def mode (plan §2+§3)”; “fix(hub): never signal a process that is no longer the node; record the node at spawn; trust only its own answer”; “fix(memory): explain a flood-stage write block in plain text (\#2284) (\#2290)”; and 45 more (refs \#2284, \#2290, \#2905, \#2104, \#1725, …).

engine/crates/xerj-hub/src · medium confidence · unverified

5 commits (0 fixes) modifying benchmarks/beir-hybrid

A change to existing behaviour in benchmarks/beir-hybrid — 5 commits, 12 files: “bench: FiQA judge row cites its raw log and the re-run spread”; “measure(beir-hybrid): stemming gate \#1059 — zero-hit 25-\>15 (FAIL), BM25 +0.0160/+0.0179 (PASS)”; “beir-hybrid: local-judge gate \#1060 measured — FAIL on quality, PASS on cost”; and 2 more (refs \#1059, \#1060, \#946, \#1091).

benchmarks/beir-hybrid · medium confidence · unverified

5 commits (0 fixes) modifying benchmarks/decisions-as-retrieval

A change to existing behaviour in benchmarks/decisions-as-retrieval — 5 commits, 29 files: “bench: commit the two gate-evidence files .gitignore '\*.md' silently dropped”; “benchmarks: FiQA (BEIR) and AG News (HF) — two public datasets added with provenance, runs and validation”; “feat(search): rerank stage with a pluggable judge provider (\#946)”; and 2 more (refs \#946, \#1057, \#1061).

benchmarks/decisions-as-retrieval · medium confidence · unverified

5 commits (0 fixes) modifying scripts/bench

A change to existing behaviour in scripts/bench — 5 commits, 10 files: “site: canonical description, synced figures, FAQPage and MCP listing manifests (\#2323)”; “bench(scripts): one-command repros for the four engine headline numbers (\#806)”; “feat(autoindex): --skills finds and indexes skill folders (\#2644) (\#2802)”; and 2 more (refs \#2323, \#806, \#2644, \#2802, \#2448).

scripts/bench · medium confidence · unverified

5 commits (0 fixes) modifying tools/benchmarks/batch-retrieval

A change to existing behaviour in tools/benchmarks/batch-retrieval — 5 commits, 7 files: “Record full-evidence batch compatibility on the default-budget release”; “Record independent Linux correctness and lifecycle receipts for batch retrieval”; “bench: record reproducible rust-vulns CLI and MCP results”; and 2 more.

tools/benchmarks/batch-retrieval · medium confidence · unverified

5 commits (2 fixes) modifying tools/benchmarks/harbor-3arm

A change to existing behaviour in tools/benchmarks/harbor-3arm — 5 commits (2 fixs), 10 files: “tools: reproducible 3-arm Harbor benchmark harness (Claude Code vs XERJ vs tuned)”; “polish(scripts): harbor-3arm fails loudly when XERJ\_BIN is unset (\#1561)”; “fix(bench): read Harbor trial rewards and account recorded attempt metrics (\#2089, from \#2070)”; and 2 more (refs \#1561, \#2089, \#2070, \#516).

tools/benchmarks/harbor-3arm · medium confidence · unverified

6 commits (0 fixes) modifying benchmarks/idle-budget

A change to existing behaviour in benchmarks/idle-budget — 6 commits, 21 files: “bench(\#874): pin THP + allocator decay off for the measured node — the runner's hugepage floor was the first CI red”; “bench(\#874): split the RSS product line from the CI gate line — 204.8 measured, 256 gated”; “bench(\#1024): idle-budget calibration after tranche 1 — 206 -\> 64 kB/idx”; and 3 more (refs \#874, \#1024, \#1025).

benchmarks/idle-budget · medium confidence · unverified

6 commits (1 fix) modifying tools/session-import

A change to existing behaviour in tools/session-import — 6 commits (1 fix), 4 files: “docs: record final Codex importer artifact verification”; “feat(sessions): import a claude-mem store with --format claude-mem (\#2874)”; “feat(cli): prepare Claude Code session text for autoindex (\#2406)”; and 3 more (refs \#2874, \#2406, \#2894).

tools/session-import · medium confidence · unverified

61 commits (6 fixes) modifying landing/answers

A change to existing behaviour in landing/answers — 61 commits (6 fixs), 194 files: “landing: keep the Day/Night choice across pages (\#2169)”; “feat(share): share links — CLI, guest reading room, tunnel (\#947)”; “feat(console): corpus home, a reader for any record, read-only guest mode (supersedes \#923) (\#945)”; and 58 more (refs \#2169, \#947, \#923, \#945, \#918, …).

landing/answers · medium confidence · unverified

7 commits (0 fixes) modifying metrics-worker

A change to existing behaviour in metrics-worker — 7 commits, 3 files: “metrics-worker: README matches the code (key-gated routes, metrics/dashboard.json)”; “metrics-worker dashboard: bar chart, tile-legend totals right-aligned, log-scale toggle, selections persist across live refresh”; “metrics-worker: make all HTTP routes private (require ?key=REFRESH\_KEY)”; and 4 more (refs \#1524, \#1531).

metrics-worker · medium confidence · unverified

729 commits (242 fixes) modifying (repo-wide)

A change to existing behaviour in (repo-wide) — 729 commits (242 fixs), 26 files: “docs: \[Unreleased\] entry for the max\_tokens MCP budget PR \#1067”; “fix(engine): delete a data stream's backing indices before recording the removal”; “changelog: cite PR \#1084 in the xerj-decide entry”; and 725 more (refs \#1067, \#1084, \#3034, \#947, \#908, …).

(repo-wide) · low confidence · unverified

75 commits (7 fixes) modifying landing/blog

A change to existing behaviour in landing/blog — 75 commits (7 fixs), 41 files: “landing: keep the Day/Night choice across pages (\#2169)”; “polish(site): rc.80 gates post lede names rc.80 again (\#1551)”; “release: cut v1.0.0-rc.81 (agent intake, Corpus Hub, \#1122 fix)”; and 71 more (refs \#2169, \#1551, \#1122, \#1371, \#2410, …).

landing/blog · medium confidence · unverified

79 commits (15 fixes) modifying scripts/seo

A change to existing behaviour in scripts/seo — 79 commits (15 fixs), 45 files: “landing: keep the Day/Night choice across pages (\#2169)”; “feat(share): share links — CLI, guest reading room, tunnel (\#947)”; “test(seo): retain offline canonical identity regressions”; and 76 more (refs \#2169, \#947, \#1371, \#2410, \#2094, …).

scripts/seo · medium confidence · unverified

81 commits (18 fixes) modifying demo/playbooks

A change to existing behaviour in demo/playbooks — 81 commits (18 fixs), 50 files: “perf(aggs): scripted\_metric sum shape → columnar .dv fast path (+730ms → 0ms)”; “bench: 81W/8L — beat 6 of the 14 (ingest c8, global agg, +4 marginal reads)”; “feat(init): install the xerj-hub skill and print the MCP registration per client (plan §9)”; and 78 more (refs \#371, \#392, \#207, \#247).

demo/playbooks · medium confidence · unverified

93 commits (0 fixes) modifying metrics

A change to existing behaviour in metrics — 93 commits, 2 files: “chore(metrics): release downloads 2026-09-19 — 2588 assets, 1448 binaries \[skip ci\]”; “chore(metrics): release downloads 2026-09-03 — 2180 assets, 1140 binaries \[skip ci\]”; “chore(metrics): release downloads 2026-09-21 — 2659 assets, 1486 binaries \[skip ci\]”; and 90 more.

metrics · medium confidence · unverified

Fixes

1 commit (1 fix) fixing benchmarks/autoindex-resilience

A fix in benchmarks/autoindex-resilience — 1 commit (1 fix), 23 files: “fix(autoindex): a refused dataset, a throttling node and an over-size catalog no longer end the run (\#934)” (refs \#934).

benchmarks/autoindex-resilience · medium confidence · unverified

1 commit (1 fix) fixing benchmarks/search-exact-phrase

A fix in benchmarks/search-exact-phrase — 1 commit (1 fix), 6 files: “fix(search): rank a file holding the exact query phrase first (\#1759)” (refs \#1759).

benchmarks/search-exact-phrase · medium confidence · unverified

1 commit (1 fix) fixing tools/cjk-judge

A fix in tools/cjk-judge — 1 commit (1 fix), 2 files: “fix(judge): score CJK evidence inside continuous text (\#2635)” (refs \#2635).

tools/cjk-judge · medium confidence · unverified

1 commit (1 fix) fixing tools/code-size-limit

A fix in tools/code-size-limit — 1 commit (1 fix), 3 files: “fix(autoindex): report exact code size-limit skips (\#2789)” (refs \#2789).

tools/code-size-limit · medium confidence · unverified

1 commit (1 fix) fixing tools/startup-diagnostics

A fix in tools/startup-diagnostics — 1 commit (1 fix), 3 files: “fix(server): report effective data directory provenance at startup (\#2414)” (refs \#2414).

tools/startup-diagnostics · medium confidence · unverified

156 commits (88 fixes) fixing engine/crates/xerj-common/src

A fix in engine/crates/xerj-common/src — 156 commits (88 fixs), 26 files: “style: clear the remaining clippy 1.97 lints (-D warnings green)”; “fix(autoindex): agg-free catalog read-back + windowed finalize refresh (\#1183)”; “fix(storage): honor wal\_sync="sync" on bulk paths + implement the wal\_batch\_ms fsync loop (RC4 W1 \#9)”; and 153 more (refs \#1183, \#9, \#2809, \#320, \#1544, …).

engine/crates/xerj-common/src · medium confidence · unverified

16 commits (9 fixes) fixing demo/usecases/autoindex

A fix in demo/usecases/autoindex — 16 commits (9 fixs), 5 files: “feat(autoindex): zero-config folder discovery + indexing as a built-in subcommand”; “fix(autoindex): scope the --fresh refusal, version the legacy migration, keep no-graph crash coverage”; “fix(autoindex): refuse unsupported corpus reconciliation”; and 13 more (refs \#274, \#2197, \#1634).

demo/usecases/autoindex · medium confidence · unverified

177 commits (94 fixes) fixing engine/crates/xerj-server/src

A fix in engine/crates/xerj-server/src — 177 commits (94 fixs), 8 files: “feat(share): share links — CLI, guest reading room, tunnel (\#947)”; “feat(hub): Corpus Hub v2 keystone — pack build/pull + def mode (plan §2+§3)”; “docs(cli): unified-engine tagline in --help and crate description”; and 174 more (refs \#947, \#1685, \#1687, \#1435, \#2284, …).

engine/crates/xerj-server/src · medium confidence · unverified

418 commits (287 fixes) fixing engine/crates/xerj-api/src

A fix in engine/crates/xerj-api/src — 418 commits (287 fixs), 29 files: “watcher: carry p\_cal beside the raw p on fire records (\#1063 landed mid-flight)”; “fix(reindex): reject source.remote with an explicit 400 instead of silently reindexing LOCAL data”; “feat(share): share links — CLI, guest reading room, tunnel (\#947)”; and 408 more (refs \#1063, \#947, \#2277, \#2022, \#202, …).

engine/crates/xerj-api/src · medium confidence · unverified

47 commits (35 fixes) fixing engine/crates/xerj-common/src/xccode

A fix in engine/crates/xerj-common/src/xccode — 47 commits (35 fixs), 8 files: “fix(xccode): only take \code\ as the passage on symbol-shaped records (\#1158)”; “feat(hub): Corpus Hub v2 keystone — pack build/pull + def mode (plan §2+§3)”; “fix(xccode): cap top-k at 2 records per source file (\#1137)”; and 44 more (refs \#1158, \#1137, \#1892, \#1908, \#1674, …).

engine/crates/xerj-common/src/xccode · medium confidence · unverified

5 commits (3 fixes) fixing engine/crates/xerj-compress

A fix in engine/crates/xerj-compress — 5 commits (3 fixs), 5 files: “fix(compress): route BitsetEnum overflow into \_\other\\_ instead of dropping docs”; “chore(lint): clear all 146 outstanding clippy + dead-code lints”; “Xerj monorepo: engine, Xerj Console, landing, demo, brief”; and 2 more (refs \#182).

engine/crates/xerj-compress · medium confidence · unverified

5 commits (4 fixes) fixing scripts/landing

A fix in scripts/landing — 5 commits (4 fixs), 7 files: “fix(playground): restore saved views without stale URL state (\#2094, from \#2071)”; “fix(playground): resolve custom dashboard templates consistently (\#2088, from \#2072)”; “fix(playground): retain query marker paint in PNG exports (\#2092, from \#2074)”; and 2 more (refs \#2094, \#2071, \#2088, \#2072, \#2092, …).

scripts/landing · medium confidence · unverified

52 commits (31 fixes) fixing engine/crates/xerj-fts

A fix in engine/crates/xerj-fts — 52 commits (31 fixs), 8 files: “perf(ingest): P2.2 no-reparse flush + ahash memtable + gated ignore\_malformed rewrite — 1M bulk 68.5k -\> 110k docs/s”; “fix(\#937): honour a declared default analyzer at flush, segment query and merge”; “docs(fts): disclose that postings skip-list acceleration is not implemented”; and 49 more (refs \#937, \#892, \#895, \#1092, \#193, …).

engine/crates/xerj-fts · medium confidence · unverified

567 commits (387 fixes) fixing engine/crates/xerj-autoindex/src

A fix in engine/crates/xerj-autoindex/src — 567 commits (387 fixs), 123 files: “fix(autoindex): re-clone a repo left half-cloned instead of skipping it (\#2566)”; “fix(search): blame the prefix only when the index is missing (\#1718)”; “watcher: carry p\_cal beside the raw p on fire records (\#1063 landed mid-flight)”; and 562 more (refs \#2566, \#1718, \#1063, \#947, \#490, …).

engine/crates/xerj-autoindex/src · medium confidence · unverified

656 commits (437 fixes) fixing engine/crates/xerj-engine/src

A fix in engine/crates/xerj-engine/src — 656 commits (437 fixs), 35 files: “perf(search): F2 — random-access pre-filtered hits via warm-slice offsets”; “perf(aggs): scripted\_metric sum shape → columnar .dv fast path (+730ms → 0ms)”; “fix(engine): delete a data stream's backing indices before recording the removal”; and 653 more (refs \#423, \#947, \#939, \#2427, \#2468, …).

engine/crates/xerj-engine/src · medium confidence · unverified

7 commits (5 fixes) fixing engine/crates/xerj-wasm

A fix in engine/crates/xerj-wasm — 7 commits (5 fixs), 5 files: “fix(fts,wasm,api): apply the \#204 sweep to the sweep's own new surface”; “fix(engine,api): fail closed on user configuration xerj cannot honour”; “chore(lint): clear all 146 outstanding clippy + dead-code lints”; and 4 more (refs \#204).

engine/crates/xerj-wasm · medium confidence · unverified

80 commits (45 fixes) fixing engine/crates/xerj-storage/src

A fix in engine/crates/xerj-storage/src — 80 commits (45 fixs), 13 files: “Cut cold first-query search latency (\#2422) (\#2443)”; “fix(storage): honor wal\_sync="sync" on bulk paths + implement the wal\_batch\_ms fsync loop (RC4 W1 \#9)”; “fix(merge): reserve merges on their on-disk footprint; drop held \_source trees (Refs \#2427) (\#2462)”; and 77 more (refs \#2422, \#2443, \#9, \#2427, \#2462, …).

engine/crates/xerj-storage/src · medium confidence · unverified

82 commits (47 fixes) fixing engine/crates/xerj-mcp/src

A fix in engine/crates/xerj-mcp/src — 82 commits (47 fixs), 9 files: “feat(memory): hybrid recall — \hybrid: true\ fuses BM25 and semantic inside /\_memory/\_recall (\#918)”; “fix(search): refuse a comma-separated --prefix instead of joining it (\#2418)”; “mcp: xerj\search defaults its index to ax-\ — the autoindex prefix (\#962)”; and 79 more (refs \#918, \#2418, \#962, \#2284, \#2290, …).

engine/crates/xerj-mcp/src · medium confidence · unverified

9 commits (6 fixes) fixing functions

A fix in functions — 9 commits (6 fixs), 5 files: “site: capture all email forms to R2 via a Pages Function”; “feat(functions): add /api/field-report HTTP fallback for agent field reports (\#1760)”; “fix(lead): reject non-string email inputs (\#2854)”; and 6 more (refs \#1760, \#2854, \#2845, \#2425).

functions · medium confidence · unverified

90 commits (63 fixes) fixing engine/crates/xerj-query

A fix in engine/crates/xerj-query — 90 commits (63 fixs), 10 files: “Convert 12 AM to hour 0 in the 12-hour (\h\/\a\) date format”; “fix(knn): thread filter+boost through top-level kNN, honor similarity cutoff, accept knn:\[...\] arrays”; “fix(highlight): expand field-name patterns in highlight.fields (\#1281)”; and 87 more (refs \#1281, \#681, \#1812, \#1816, \#1947, …).

engine/crates/xerj-query · medium confidence · unverified

Test coverage

10 commits adding/updating tests in engine/crates/xerj-cluster/tests; 18 commits adding/updating tests in engine/crates/xerj-rerank/tests; 187 commits adding/updating tests in engine/crates/xerj-api/tests; 19 commits adding/updating tests in engine/crates/xerj-autoindex/tests; 19 commits adding/updating tests in engine/crates/xerj-hub/tests; 2 commits adding/updating tests in engine/crates/xerj-storage/tests; 22 commits adding/updating tests in engine/crates/xerj-mcp/tests; 223 commits adding/updating tests in engine/crates/xerj-engine/tests; 24 commits adding/updating tests in engine/crates/xerj-console-api/tests; 27 commits adding/updating tests in engine/tests/es-compat-yaml; 3 commits adding/updating tests in engine/tests; 33 commits adding/updating tests in engine/crates/xerj-server/tests; 5 commits adding/updating tests in engine/crates/xerj-common/tests; 6 commits adding/updating tests in engine/crates/xerj-autoindex/tests/fixtures.

Dependencies

182 commits updating dependencies (25 manifests)

A dependency / build maintenance change in (dependencies) — 182 commits (42 fixs), 25 files: “feat(share): share links — CLI, guest reading room, tunnel (\#947)”; “fix(http2): raise the small-DATA-frame budget with the window (h2 0.4.19) (\#908)”; “release: cut v1.0.0-rc.81 (agent intake, Corpus Hub, \#1122 fix)”; and 178 more (refs \#947, \#908, \#1122, \#2268, \#1869, …).

(dependencies) · high confidence · unverified

Written by cca.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.

How this codebase got here

Baseline

  • First survey — no prior run to compare against. CAI 41.6.

Lenses

  • Code Health 19.1
  • Architecture 93.0
  • Maturity 71.2
  • Readiness 62.3
  • Security 54.2
  • Event-Driven 71.9
  • Event Sourcing 100.0
  • Accessibility 61.2
  • Performance 70.7

Changes since last survey

  • 300 commits — 203 feature/other, 97 fixes

By area

  • engine/crates — 100 commits
  • (repo) — 59 commits
  • (root) — 57 commits
  • user-feedback/16-agent-field-reports — 16 commits
  • metrics/release-downloads.jsonl — 12 commits
  • landing/answers — 11 commits
  • landing/blog — 7 commits
  • landing/docs — 4 commits
  • landing/index.html — 3 commits
  • landing/xerj-query-1.html — 3 commits
  • scripts/seo — 3 commits
  • .github/scripts — 2 commits
  • docs/RERANK.md — 2 commits
  • engine/Cargo.lock — 2 commits
  • .github/workflows — 1 commit
  • benchmarks/query-1-2026-11 — 1 commit
  • content/visuals — 1 commit
  • docs/FACTS_API.md — 1 commit
  • docs/LIVE_REINDEXING.md — 1 commit
  • docs/examples — 1 commit

Notable commits

  • fix: Compose alias reindex preflight and buffered Boolean scoring fixes
  • fix: Merge pull request #3023 from xerj-org/fix/claims-corrections-20261106
  • fix: Merge pull request #3028 from xerj-org/fix/3021-facts-startup
  • fix: Merge pull request #3029 from xerj-org/fix/3005-profile-honest
  • fix: Merge pull request #3030 from xerj-org/fix/3017-hybrid-tiebreak
  • fix: Merge pull request #3031 from xerj-org/fix/functions-get-latlong-flake
  • fix: Merge pull request #3037 from xerj-org/fix/3036-bulk-put
  • fix: agents: XERJ-Query in llms.txt, AGENTS.md and README; field-report ask fixed
  • fix: autoindex: index INSERT-only seed migrations instead of junking them (Fixes #2877) (#2878)
  • fix: ci(fuzz): pin nightly-2026-10-30 (rustc 2026-10-31 E0391 regression breaks every PR's fuzz job) (#2832)
  • fix: fix(aggs): bound hdr number_of_significant_value_digits; one request hung a search thread (#2651)
  • fix: fix(aggs): execute weighted_avg and fail loud on unsupported shapes (#2805)
  • fix: fix(aggs): keep composite missing order consistent across pages (#2944)
  • fix: fix(ai): accept embedding responses without a per-item index
  • fix: fix(api): _reindex applies source filtering and refuses scripts (#2719)
  • fix: fix(api): _reindex honours op_type, conflicts and max_docs (#2653)
  • fix: fix(api): _search profile reports measured phase timings, not placeholders (#3005)
  • fix: fix(api): delete a one-page _delete_by_query match set without a flush (#2782)
  • fix: fix(api): enforce index.max_terms_count on terms queries (#2936)
  • fix: fix(api): enforce the search.max_keep_alive cluster setting (#2937)
  • …and 280 more

Architecture

  • 0 containers · 1 bounded contexts · 0 dependency edges (baseline)

Written by cca.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.

Survey your own repository

xerj-org/xerj was measured the same way every project in this corpus was: the same rubric, at a pinned commit, with the result published in full. Point a surveyor at a repository you know and see whether you agree with it.

About this page

  • The score is its most recent published measurement, taken on 17 November 2026 at a pinned commit. It is not a live figure and does not change until the project is measured again.
  • Measured at commit 50e4ffded5518b1822cfac5450fa34d3829ab8ab — the exact code this score is about.
  • Scored under rubric-2026.11.4 — the same rubric and the same method as every other entry in this index.
  • Measured by Canine Code Analyzer (cca.canine.dev) using cca measure-2026.11.85+preprod-eba6548ac50f.