{"full_name":"xberg-io/xberg","owner":"xberg-io","name":"xberg","shard":"1","url":"https://github.com/xberg-io/xberg","description":"The fast, precise document-intelligence engine for every language: text, tables and metadata from 100+ formats, with OCR, transcription and code intelligence.","homepage":"https://docs.xberg.io","topics":["text-extraction","document-intelligence","metadata-extraction","pdf-extraction","python","rag","table-extraction","tesseract","ffi","golang","java","node","ruby","rust","wasm"],"tag_pages":["ffi","golang","java","node","python","rag","ruby","rust","tesseract","text-extraction","wasm"],"language":"Rust","license":"MIT","avatar":"https://avatars.githubusercontent.com/u/241328462?v=4","created_at":"2025-01-31T21:50:02Z","pushed_at":"2026-11-15T14:44:52Z","metrics":{"stars":9590,"forks":617,"watchers":32,"issues":1631,"prs":1498,"releases":127,"issues_open":23,"issues_closed":1608,"prs_open":1,"prs_closed":345,"prs_merged":1152,"discussions":32},"growth":{"day":{"stars":5,"forks":1,"base_day":"2026-11-15"}},"history":{"dates":["2026-11-10","2026-11-11","2026-11-12","2026-11-13","2026-11-14","2026-11-15","2026-11-16"],"stars":[9569,9576,9581,9581,9583,9585,9590],"forks":[613,614,614,615,615,616,617],"watchers":[null,32,32,32,32,32,32],"issues_open":[null,66,73,74,29,18,23],"prs_open":[null,19,19,14,1,1,1]},"boards":{},"same_owner":[],"issues":[{"number":2332,"title":"fix(tables): preserve HTML headers, spans and table boundaries","comments":16,"created_at":"2026-10-30T06:54:54Z"},{"number":1567,"title":"chore(lint): pay down the size and complexity baseline","comments":14,"created_at":"2026-09-05T15:12:09Z"},{"number":2502,"title":"fix(ocr): tall-scan strip recovery covers only a narrow set of scans","comments":12,"created_at":"2026-11-05T01:48:32Z"},{"number":2331,"title":"fix(ocr): preserve math and scanned text; separate headers and footers","comments":7,"created_at":"2026-10-30T06:54:52Z"},{"number":2503,"title":"fix(ocr): table rows still split across lines after the exact-row claim","comments":5,"created_at":"2026-11-05T01:48:35Z"},{"number":2335,"title":"feat(onnx): add CPU classification and captioning backends","comments":3,"created_at":"2026-10-30T16:46:38Z"},{"number":2334,"title":"perf(onnx): qualify CPU classification and captioning models","comments":3,"created_at":"2026-10-30T16:46:32Z"},{"number":2336,"title":"feat(models): publish pinned ONNX artifacts and SDK inventory","comments":2,"created_at":"2026-10-30T16:46:57Z"}],"discussions":[{"number":1872,"title":"Bounded owner-run PDF extraction: adapter approach and research trial","comments":1,"category":"General","created_at":"2026-09-26T23:05:29Z"},{"number":1274,"title":"Option to merge image descriptions into content for single-field consumers (RAG pipelines)","comments":0,"category":"Ideas","created_at":"2026-07-19T14:23:09Z"},{"number":1200,"title":"semantic versioning and maintenance of v4","comments":3,"category":"Q&A","created_at":"2026-07-05T16:07:56Z"},{"number":1184,"title":"Kreuzberg v4.9.9 vs Xberg + patch plan for OpenSSL CVE","comments":2,"category":"Q&A","created_at":"2026-06-30T16:06:40Z"},{"number":1119,"title":"lilbee: a single-executable local search engine you can talk to. It runs the models, crawls the web into your library, and reads 90+ document formats with Kreuzberg","comments":0,"category":"Show and tell","created_at":"2026-06-14T20:50:48Z"},{"number":952,"title":"Release frequency question","comments":2,"category":"Q&A","created_at":"2026-05-13T17:38:26Z"},{"number":842,"title":"License Question","comments":1,"category":"Q&A","created_at":"2026-04-29T05:38:18Z"},{"number":737,"title":"Preserve links in markdown for images?","comments":0,"category":"General","created_at":"2026-04-17T08:26:18Z"}],"latest_release":{"name":"v1.4.0","tag":"v1.4.0","published_at":"2026-11-13T15:42:44Z","body":"## Highlights\n\n- Faster PDF extraction when only a few glyphs need OCR, with a full-resolution retry when needed.\n- Tesseract engine reuse for automatic script retries reduces repeated model loading.\n- Better preservation of native PDF text and exact page byte ranges when OCR falls back.\n- An experimental PP-DocLayout-V3 option for ONNX Runtime builds and explicit LaTeX annotations for Markdown, HTML, and Djot.\n\n[Read the full v1.4.0 changelog](https://github.com/xberg-io/xberg/blob/v1.4.0/CHANGELOG.md#unreleased).\n","is_prerelease":false},"readme":true,"readme_zh":false,"news":[],"stale":false,"snapshot_at":"2026-11-16T09:23:14Z","history_start":"2026-09-06"}