CodeNib: A Multi-View Data System for Serving Repository Context to Coding Agents

2026-07-28

AI Summary

This work addresses persistent challenges faced by coding agents in evolving code repositories—namely redundant retrieval, fragmented context, and opaque computational costs. To overcome these issues, the authors propose a multi-view context service that constructs reusable lexical, dense, and structured views for each commit, unifying support for code search, symbol navigation, and bounded context provision. By explicitly defining operation-level validity boundaries, the system enables cross-edit view persistence and efficient incremental updates. Experimental results demonstrate that, compared to full reconstruction, graph and vector index updates achieve speedups of 8.7× and 25.4×, respectively; static navigation latency is reduced to one-fourth of real-time serving requirements; and context-aware strategies cut trajectory token consumption by 50%–87%.

Abstract

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. Across 100 snapshots, we map quality-cost frontiers across the repository-context lifecycle. When outputs match an independent rebuild, graph and vector updates are $8.7\times$ and $25.4\times$ 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 $4.7\times$. 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.

Problem

• coding agents
• repository context
• context retention
• disconnected indexes
• lifecycle costs

Innovation

• multi-view indexing
• repository context serving
• coding agents
• structural code representation
• efficient context retrieval