8000 site+readme: XERJ Query-1 up front (README hero, home-page model + install sections) by xerj-org · Pull Request #2978 · xerj-org/xerj · GitHub
[go: up one dir, main page]

Skip to content

site+readme: XERJ Query-1 up front (README hero, home-page model + install sections) - #2978

Merged
xerj-org merged 3 commits into
mainfrom
site/front-model-first
Nov 5, 2026
Merged

xerj-org merged 3 commits into
mainfrom
site/front-model-first

Conversation

@xerj-org
@xerj-org xerj-org commented Nov 5, 2026 •
Copy link
Copy Markdown
Owner

XERJ Query-1 up front on the README and the home page.

Home page (landing/index.html)

  • Hero names XERJ Query-1: a search engine with a neurosymbolic model inside; your question becomes a short search program the engine runs over your data, every answer points at its record. Sub-line: no embeddings, no GPU, nine times fewer wrong answers than RAG in our test, hundreds of times cheaper per question; the same binary does full-text and hybrid search over files, folders and S3 and speaks the Elasticsearch API.
  • One agent prompt for the broad use case (index the data, search it, write its key facts into a facts namespace, ask one question; consent step; hand-off of hits, answer, program and evidence), then the shell path curl -fsSL https://xerj.org/get | sh -s -- --with-model with the start line and a verify line with expected output, and a compact no-shell row (HTTP and MCP).
  • The six measured vs-cards, then "What else XERJ does": search your files, agent memory and brains, reference coding (the former hero prompt, kept verbatim with its own copy button), Elasticsearch-compatible API.
  • Follows the llms.txt study rules: three entry surfaces, verify lines, short first screen, prompt under 300 characters, no obligation language.

README: badges on top, hero image, six-row comparison, limits, 3-command install with the model; the previous body follows under "The engine underneath".

Also: the XERJ Query-1 blog post embeds its hero cover; /xerj-query-1 links to the post.

Checks: SEO writers + every --check, seo_lint 0 gate violations, landing constants guard, test_verify_agent_discovery, test_llms_field_report, test_article_schema, test_urlmap.

🤖 Generated with Claude Code

…install sections)

Owner request: put the model and its value on the first screen of the GitHub
README and of xerj.org, then show how to install and use it with the model.

README.md: title, then the badge block (adds an "XERJ Query-1, new in rc.101"
badge; Trendshift and language links kept), then a 1600x800 hero image
(docs/media/xerj-query-1-hero.png, rendered from HTML in the site's night
palette), a one-line pitch, a link row (Query-1 page, blog, release,
llms.txt, docs, donate), and "XERJ Query-1, new in rc.101": what it does in
plain words, a six-row measured comparison table with the scope of each
number, the off-by-default and not-yet limits, and a three-command install
with verify lines and expected output. The existing body follows unchanged
under "The engine underneath". Order follows the top-starred AI repos
(ollama, qdrant, meilisearch, chroma, llama_index, DeepSeek-V3): pitch,
hero, quickstart, numbers, then features and docs.

landing/index.html: a hero line pointing at the model, then two sections
after the hero: "Ask a question. It writes the program." (plain
explanation, the six ours-vs-theirs cards shared with the blog and
/xerj-query-1, limits) and "Install it with the model" with three entry
paths per the llms.txt study rules: an agent prompt, a shell path, an
HTTP-only path and an MCP path, each ending in a verify line with the
expected output. The existing reference-coding hero prompt is unchanged.

Numbers come from SEVEN_CARDS (blog v13): 120 ms vs 2,700-3,300 ms on
namespaces up to 8,000 facts; 1 vs 9-15 wrong of 160; $42 vs
$13,400-$694,000 per million questions; 17 min vs 62-97 h for 10 GB;
0 s re-index; $468 vs $65,747 per month.

Checks: fix_heads/fix_links/gen_sitemap/mk_og_card/build_articles/set_nav/
build_skills --check ok; landing-constants-guard ok; test_llms_field_report,
test_verify_agent_discovery, test_article_schema, test_urlmap ok. seo_lint
flags /xerj-query-1 and /blog/xerj-query-ai-model until PRs #2972 and #2976
land. README strings asserted by readme_data_dir.rs are unchanged.
…uery-1

landing/blog/xerj-query-ai-model.html embeds its committed hero cover
(blog/covers/xerj-query-ai-model-hero.webp, 1600x640) under the crumb the
same way the other posts do, and loads blog.css?v=4 like them so the
blog-cover rule applies. landing/xerj-query-1.html: the blog link that
waited for publish is now live text. SEO writers re-run; all --check
modes pass, seo_lint 0 gate violations.
…t on top

The hero now names XERJ Query-1 and what it does (a search program the
engine runs over your data, every answer pointing at its record), then one
agent prompt for the broad use case: index the data, search it, write its
key facts into a facts namespace and ask a question, with a consent step
and a hand-off of hits, answer, program and evidence. Under it the shell
path with --with-model, the start line and a verify line with expected
output, plus a compact no-shell row (HTTP and MCP), per the llms.txt study
rules (three entry surfaces, verify lines, short first screen, prompt
under 300 characters, no obligation language).

The six vs-cards follow, then a new What else XERJ does section with four
tiles: search your files (autoindex), agent memory and brains, reference
coding (the former hero prompt, kept verbatim with its own copy button),
and the Elasticsearch-compatible API. The separate Install-it-with-the-
model section is folded into the hero. Detail sections get anchors.

Checks: SEO writers + every --check, seo_lint 0 gate violations, landing
constants guard, test_verify_agent_discovery, test_llms_field_report,
test_article_schema, test_urlmap all pass.
@xerj-org
xerj-org merged commit 7458a22 into main Nov 5, 2026
27 of 29 checks passed
YYW0228 pushed a commit to YYW0228/xerj that referenced this pull request Nov 7, 2026
The public half of the v1.0.0-rc.102 cut. The release tag goes on this
PR's merge commit, so the notes and the tree they describe are one commit.

- CHANGELOG.md: [Unreleased] cut over to [1.0.0-rc.102] - 2026-11-06.
  Every PR merged since v1.0.0-rc.101 is cited (xerj-org#2969 xerj-org#2982 xerj-org#2983 xerj-org#2984
  xerj-org#2985 xerj-org#2988 xerj-org#3023 xerj-org#3028 xerj-org#3029 xerj-org#3030) or exempted with its reason (site,
  blog and donation pages: xerj-org#2971 xerj-org#2972 xerj-org#2976 xerj-org#2978 xerj-org#2980 xerj-org#2990 xerj-org#3032;
  CI test-only: xerj-org#3031). The release-notes gate passes locally against
  origin/main 34a2f24.
- The XERJ Query-1 numbers in the summary trace to the R8 evaluation
  (private harness, runs/r8_2026-11-06/R8_RESULT.md): Score v2 on the four
  development families, 1,600 items, seed 47 90.29 -> 94.69 (+4.40,
  95% interval +3.32 to +5.64), seeds 11/29 +3.95/+4.17; stated as
  in-sample because R8 was chosen on those items; wrong answers 9 -> 10
  of 1,600 (s47), 12 -> 15 (s11); held-out fresh_v1 Net 0.3333 -> 0.3481.
  The 1,600/1,600 engine-vs-Python parity was re-run on the merged rc.102
  build tree.
- Upgrade notes say plainly that the rc.101 promise to land the storage
  and ingest lanes on public main before rc.102 was not kept, and that the
  rc.102 bundle needs an rc.102 binary.
- Fixed lists the rc.101 aarch64 Linux assets built with a 4 KiB jemalloc
  page (checked in the rc.101 build tree: "checking LG_PAGE... 12" for
  both aarch64 targets), which xerj-org#1663 had fixed in release.yml; rc.102's
  native aarch64 builds set JEMALLOC_SYS_WITH_LG_PAGE=16.
- ROADMAP.md: review line names v1.0.0-rc.102 (the
  docs_capability_lists check); the open-defects shortlist re-verified
  live: xerj-org#1122 and xerj-org#1170 open.
- engine/Cargo.toml workspace version 1.0.0-rc.102, Cargo.lock restamped
  (19 workspace packages, no dependency moved).
- Landing footer stamps V1.0.0-RC.101 -> V1.0.0-RC.102 (59 pages) and the
  nav's fallback latest-tag label; landing-constants-guard passes.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant

0