ctxfeed v0.6.0

Context-feeding layer for coding agents (Claude Code, Cursor) — shards a repo into model-sized context windows with a persistent cache so repeat ingestions skip unchanged files and report an honest cache-hit rate.

ctxfeed v0.5.0

Local MCP project-context backend that shards a repo into GLM-5.2's 1M-token window

ctxfeed v0.4.0

ctxfeed is a local MCP project-context backend that shards a whole repo into GLM-5.2's 1M-token window with cache-aware ingest ordering

ctxfeed v0.3.0

ctxfeed is a local MCP project-context backend that shards a whole repo into GLM-5.2's 1M-token window with cache-aware ingest ordering

ctxfeed v0.2.0

ctxfeed is a local MCP project-context backend that shards a whole repo into GLM-5.2's 1M-token window with cache-aware ingest ordering

ctxfeed v0.1.0

ctxfeed is a local MCP project-context backend that shards a whole repo into GLM-5.2's 1M-token window with cache-aware ingest ordering, so coding agents like Claude Code and Cursor query 1000+ files in one call — below Opus per-token cost, past ChatGPT's 40-file cap. Rides the Kimi K3 / DeepSeek V4 long-context wave.