{"version":1,"articles":[{"slug":"grok-4-6-vs-4-7","title":"Grok 4.7 uses more tokens per task than 4.6 — and scores higher","publishedAt":"2026-09-22","modifiedAt":"2026-09-22","url":"https://codegraff.com/blog/grok-4-6-vs-4-7","json":"https://codegraff.com/blog/grok-4-6-vs-4-7/article.json","markdown":"https://codegraff.com/blog/grok-4-6-vs-4-7/markdown"},{"slug":"codedb-with-graff","title":"How CodeDB works with Graff and your coding agent","publishedAt":"2026-09-11","modifiedAt":"2026-09-12","url":"https://codegraff.com/blog/codedb-with-graff","json":"https://codegraff.com/blog/codedb-with-graff/article.json","markdown":"https://codegraff.com/blog/codedb-with-graff/markdown"},{"slug":"graff-frontier-harness-evals","title":"Graff on FrontierHarness: coding agent benchmarks, costs and comparisons","publishedAt":"2026-09-06","modifiedAt":"2026-09-06","url":"https://codegraff.com/blog/graff-frontier-harness-evals","json":"https://codegraff.com/blog/graff-frontier-harness-evals/article.json","markdown":"https://codegraff.com/blog/graff-frontier-harness-evals/markdown"},{"slug":"graff-rlm-token-efficiency","title":"Graff's RLM runtime: how 135k input tokens became 25.3k","publishedAt":"2026-08-28","modifiedAt":"2026-08-28","url":"https://codegraff.com/blog/graff-rlm-token-efficiency"},{"slug":"codedb-retrieval-experiments","title":"We tested four ways to improve CodeDB retrieval. One earned its way in.","publishedAt":"2026-08-05","modifiedAt":"2026-08-05","url":"https://codegraff.com/blog/codedb-retrieval-experiments"},{"slug":"codedb-pro-0-2-14","title":"CodeDB Pro v0.2.14: the 50 ms wait that made fast search look slow","publishedAt":"2026-07-15","modifiedAt":"2026-07-15","url":"https://codegraff.com/blog/codedb-pro-0-2-14"},{"slug":"codedb-pro-0-2-12","title":"CodeDB Pro v0.2.12: fail-closed edits and a faster MCP response path","publishedAt":"2026-07-11","modifiedAt":"2026-07-11","url":"https://codegraff.com/blog/codedb-pro-0-2-12"},{"slug":"codedb-pro","title":"Muonry is now CodeDB Pro: same daemon, a name that says what it is","publishedAt":"2026-06-15","modifiedAt":"2026-06-15","url":"https://codegraff.com/blog/codedb-pro"},{"slug":"muonry-0-2-7","title":"muonry v0.2.7: 89× smaller edit responses, and reads that flag smuggled text","publishedAt":"2026-06-11","modifiedAt":"2026-06-11","url":"https://codegraff.com/blog/muonry-0-2-7"},{"slug":"codedb-0-2-5824","title":"codedb v0.2.5824: 3× faster cold start, and a CLI as fast as MCP","publishedAt":"2026-06-05","modifiedAt":"2026-06-05","url":"https://codegraff.com/blog/codedb-0-2-5824"},{"slug":"codedb-0-2-5818","title":"codedb v0.2.5818: SOTA retrieval, and the numbers to prove it","publishedAt":"2026-05-25","modifiedAt":"2026-05-25","url":"https://codegraff.com/blog/codedb-0-2-5818"},{"slug":"codedb-swe-bench-lite","title":"codedb on SWE-bench Lite: six backends, deterministic oracle","publishedAt":"2026-05-22","modifiedAt":"2026-05-22","url":"https://codegraff.com/blog/codedb-swe-bench-lite"},{"slug":"codegraff-0-2-4","title":"Codegraff v0.2.4: from one replace surface to a daemon that doesn't hang","publishedAt":"2026-05-21","modifiedAt":"2026-05-21","url":"https://codegraff.com/blog/codegraff-0-2-4"},{"slug":"muonry-0-1-15","title":"muonry v0.1.15: 21 silent failures fixed, and a token-efficiency benchmark","publishedAt":"2026-05-08","modifiedAt":"2026-05-08","url":"https://codegraff.com/blog/muonry-0-1-15"},{"slug":"zig-0-16-migration","title":"What Zig 0.16's new stdlib did to our hot paths","publishedAt":"2026-04-17","modifiedAt":"2026-04-17","url":"https://codegraff.com/blog/zig-0-16-migration"},{"slug":"muonry-0-1-13-performance","title":"muonry v0.1.13: 20% faster reads, even with safety scanning on every file","publishedAt":"2026-04-16","modifiedAt":"2026-04-16","url":"https://codegraff.com/blog/muonry-0-1-13-performance"},{"slug":"codedb-0-2-572","title":"codedb v0.2.572: 10x faster indexing, 83% less memory","publishedAt":"2026-04-15","modifiedAt":"2026-04-15","url":"https://codegraff.com/blog/codedb-0-2-572"},{"slug":"codedb-0-2-53-to-0-2-56","title":"codedb's 70x warm-start fix","publishedAt":"2026-04-09","modifiedAt":"2026-04-09","url":"https://codegraff.com/blog/codedb-0-2-53-to-0-2-56"},{"slug":"codedb-code-intelligence","title":"Why codedb feels instant for agents","publishedAt":"2026-04-01","modifiedAt":"2026-04-01","url":"https://codegraff.com/blog/codedb-code-intelligence"},{"slug":"muonry-edit-efficiency","title":"78% fewer tokens: how we made muonry edits 4x more efficient","publishedAt":"2026-03-14","modifiedAt":"2026-03-14","url":"https://codegraff.com/blog/muonry-edit-efficiency"}]}