Coding agents keep rediscovering the same repository. CodeNib treats repo context as a data-systems problem: build lexical, dense, and structural views once per commit, maintain each through its own incremental path, and serve them as bounded agent context.</p>\n<p>🚀 8.7×/25.4× faster graph/vector updates than rebuilding (only counted where outputs match an independent rebuild)<br>🧭 4.7× lower median latency than a live LSP on compatible navigation requests<br>🪙 50–87% fewer trajectory tokens than grep/read at matched localization quality, across 5 agent models</p>\n<p>Code &amp; artifacts: github.com/sysevol-ai/CodeNib</p>\n","updatedAt":"2026-07-29T10:00:42.972Z","author":{"_id":"652622617fa0bebfbf0e5dd1","avatarUrl":"/avatars/765ad731868431b80ea751d666998714.svg","fullname":"Zhongming Yu","name":"fishmingyu","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":1,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.8679980039596558},"editors":["fishmingyu"],"editorAvatarUrls":["/avatars/765ad731868431b80ea751d666998714.svg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2607.25431","authors":[{"_id":"6a69ce1d9d3a1231d492ba35","name":"Zhongming Yu","hidden":false},{"_id":"6a69ce1d9d3a1231d492ba36","name":"Hengjia Yu","hidden":false},{"_id":"6a69ce1d9d3a1231d492ba37","name":"Boqin Yuan","hidden":false},{"_id":"6a69ce1d9d3a1231d492ba38","name":"Shuting Zhao","hidden":false},{"_id":"6a69ce1d9d3a1231d492ba39","name":"Yizhao Chen","hidden":false},{"_id":"6a69ce1d9d3a1231d492ba3a","name":"Aryan Dokania","hidden":false},{"_id":"6a69ce1d9d3a1231d492ba3b","name":"Mihir Jagtap","hidden":false},{"_id":"6a69ce1d9d3a1231d492ba3c","name":"Jiayu Chang","hidden":false},{"_id":"6a69ce1d9d3a1231d492ba3d","name":"Yitong Ma","hidden":false},{"_id":"6a69ce1d9d3a1231d492ba3e","user":{"_id":"654b5754fabd2cc668763f86","avatarUrl":"/avatars/c4cbaa9b3af35783b0a163bbba49839a.svg","isPro":false,"fullname":"Yash Jayswal","user":"NameUrSis","type":"user","name":"NameUrSis"},"name":"Yash Jayswal","status":"claimed_verified","statusLastChangedAt":"2026-07-29T16:45:04.799Z","hidden":false},{"_id":"6a69ce1d9d3a1231d492ba3f","name":"Wentao Ni","hidden":false},{"_id":"6a69ce1d9d3a1231d492ba40","name":"Hejia Zhang","hidden":false},{"_id":"6a69ce1d9d3a1231d492ba41","name":"Zhaoling Chen","hidden":false},{"_id":"6a69ce1d9d3a1231d492ba42","name":"Gangda Deng","hidden":false},{"_id":"6a69ce1d9d3a1231d492ba43","name":"Jishen Zhao","hidden":false}],"mediaUrls":["https://cdn-uploads.huggingface.co/production/uploads/652622617fa0bebfbf0e5dd1/1jvWku-w5Cv__NLuG8s_-.mp4"],"publishedAt":"2026-07-28T00:00:00.000Z","submittedOnDailyAt":"2026-07-29T00:00:00.000Z","title":"CodeNib: A Multi-View Data System for Serving Repository Context to Coding Agents","submittedOnDailyBy":{"_id":"652622617fa0bebfbf0e5dd1","avatarUrl":"/avatars/765ad731868431b80ea751d666998714.svg","isPro":false,"fullname":"Zhongming Yu","user":"fishmingyu","type":"user","name":"fishmingyu"},"summary":"Coding agents repeatedly search, navigate, and retain context from evolving repositories, but disconnected indexes, language servers, and task-local histories force repeated discovery and obscure lifecycle costs. CodeNib builds reusable lexical, dense, and structural views per repository commit, maps outputs to repository-relative source ranges, maintains selected views across edits, and serves ranked search, symbol navigation, and bounded context through one runtime.\n Across 100 snapshots, we map quality-cost frontiers across the repository-context lifecycle. When outputs match an independent rebuild, graph and vector updates are 8.7times and 25.4times faster at the median. On the static-navigation subset matching normalized live-server locations (63% of 1,000 requests), the median per-request live/static latency ratio is 4.7times. Across five models, selected context policies preserve localization with 50--87% fewer trajectory tokens than paired grep/read. Together, these results support multi-view repository-context serving with explicit, operation-specific validity boundaries.","upvotes":22,"discussionId":"6a69ce1d9d3a1231d492ba44","projectPage":"https://codenib.ai","githubRepo":"https://github.com/sysevol-ai/CodeNib","githubRepoAddedBy":"user","githubStars":7,"organization":{"_id":"6a0e7eccbf416c175d0e849e","name":"sysevol-ai","fullname":"SysEvol AI Research","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/652622617fa0bebfbf0e5dd1/vESwUVoVpA7G0KgEsrib1.png"}},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"652622617fa0bebfbf0e5dd1","avatarUrl":"/avatars/765ad731868431b80ea751d666998714.svg","isPro":false,"fullname":"Zhongming Yu","user":"fishmingyu","type":"user"},{"_id":"698f03062318dbb8ef930188","avatarUrl":"/avatars/da98add3194e74f4e34faabdec6dba39.svg","isPro":true,"fullname":"hyd2apse","user":"hyd2apse","type":"user"},{"_id":"688cae87ef8536018361d635","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/no-auth/321u9EioGGWK5jKA-anZj.png","isPro":false,"fullname":"Siriux Yu","user":"Siiriuxxx","type":"user"},{"_id":"64ed861f254de0e729dba8f1","avatarUrl":"/avatars/3cca092280a8e34826657a5ba300ca68.svg","isPro":false,"fullname":"C.F.A","user":"codefuse-admin","type":"user"},{"_id":"6a69d534221f3cc2d9492240","avatarUrl":"/avatars/660db61e5d2b4631eb2c9358fc07b23a.svg","isPro":false,"fullname":"Jiashen Ren","user":"GaAs9000","type":"user"},{"_id":"6a69d722c6d079e5e6bc9b47","avatarUrl":"/avatars/c429b3ee5f27b85839292e73efaa4752.svg","isPro":false,"fullname":"Wu Boxiang","user":"Hal227","type":"user"},{"_id":"6a69d839c6d079e5e6bca8fa","avatarUrl":"/avatars/744c263ac52ab41fc63192488c322d25.svg","isPro":false,"fullname":"geehyun jung","user":"geehyunj","type":"user"},{"_id":"620783f24e28382272337ba4","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/620783f24e28382272337ba4/zkUveQPNiDfYjgGhuFErj.jpeg","isPro":false,"fullname":"GuoLiangTang","user":"Tommy930","type":"user"},{"_id":"6a69e006b7d15c53d5f84735","avatarUrl":"/avatars/4a1cb85bddd6af003af8f4abc4fae5de.svg","isPro":false,"fullname":"James Auster","user":"Auster001","type":"user"},{"_id":"6a69e10a169a610b5d14038c","avatarUrl":"/avatars/6d8c71d3a01b04ab5de292a3434e70f9.svg","isPro":false,"fullname":"haoyu","user":"haoyuhua333","type":"user"},{"_id":"6a69e41ca723582772ddfe92","avatarUrl":"/avatars/cf68ea5820c6eea466627afb93e0934c.svg","isPro":false,"fullname":"Andrea Jose","user":"andreaneenu","type":"user"},{"_id":"6a69e7693b31270b8b832d72","avatarUrl":"/avatars/a5147f801f4d3d38ca77ffd97f4387d0.svg","isPro":false,"fullname":"Sylvia Yao","user":"Sylvia0728","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"organization":{"_id":"6a0e7eccbf416c175d0e849e","name":"sysevol-ai","fullname":"SysEvol AI Research","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/652622617fa0bebfbf0e5dd1/vESwUVoVpA7G0KgEsrib1.png"},"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2607/2607.25431.md","query":{}}">

Get this paper in your agent:

Don't have the latest CLI?

Cite arxiv.org/abs/2607.25431 in a model README.md to link it from this page.

Cite arxiv.org/abs/2607.25431 in a Space README.md to link it from this page.

Add this paper to a collection to link it from this page.