Repository Issue Activity (beta)

#556 [Bug]: JSON text fallback raises TypeError when malformed input exceeds chunk_chars

#506 [Feature]: Add first-class memory subject scoping

#386 feat: support horizontal scaling / cluster deployment for production

#523 [Bug]: Embedding backfill/rebuild dies instantly on 429 rate-limit (no retry/backoff) and leaves vector search permanently disabled after failure

#524 [Bug]: Persisted chunk store is not invalidated when the chunker changes — stale "full-outline" chunks survive upgrades and are never re-chunked

#378 feat(memory): scope procedural memory by project

#464 Make embedding health check failures recoverable

#463 SearchStep and _ToolContextDedupMixin share the class constant SEARCH_SEEN_KEY with incompatible value shapes

#429 Chinese frontmatter description not searchable via BM25; `stat`/`frontmatter_read`/`frontmatter_update` do not auto-append `.md` suffix

#427 [Feature] 支持项目级记忆隔离，并区分通用记忆和项目记忆

#350 feat(plugin): add ReMe integration for Codex

#393 Feature Request: Add optional "forgotten memory" detection for stale TODOs in daily notes

#388 fix(file_io): ReadStep small-file path over-counts total lines by 1 for trailing-newline files

#383 Auto Resource Data Connector framework (with daily_paper as the first connector)

#374 session memory management

#352 feat(agent-wrapper): add Codex backend

#368 新版本不支持根据用户查询了吗

#364 咨询reme v0.4.x版本功能

#373 [Question] What led to the decision to make file-native memory the primary paradigm and demote vector search to an opt-in layer in ReMe v0.4?

#376 bug: MCP tool responses lose structured metadata — only human-readable text is returned

#379 fix: make metadata-backed step answers complete for LLM consumers

#347 feat(plugin): add ReMe integration for OpenClaw

#348 feat(plugin): add ReMe integration for Hermes Agent

#360 perf(indexing): Bound auto-resource ingestion and reduce InitChangesStep startup memory

#353 benchmark: add LoCoMo evaluation