Comunidad Guías…

Ayuda y problemas de mem0

Pregunta a la comunidad o invita a KiBot con contexto del repositorio.

Publicar un problema

No publiques contraseñas, tokens ni claves privadas. HostingByte aplica una redacción básica de secretos evidentes antes de guardar el texto.

Problemas en HostingByte

Todavía no hay problemas publicados.

Issues abiertas upstream

#7811 FAISS: a payload with a datetime or numpy value truncates the docstore JSON, and every memory in the collection is lost on restart · 5 comentarios #7891 Coding-agent plugins: oversized memory hides later recall results and marks them as injected · 3 comentarios #7865 Memory.add(): report why a fact was not stored (skipped duplicate / failed insert) instead of silently omitting it · 3 comentarios #7889 bug(vector_stores/faiss): insert() with an existing id adds a second vector instead of replacing it, so search returns duplicates and delete leaves an orphan · 3 comentarios #5352 Workaround: Solving time-blindness, semantic conflicts, and missing CRUD in Mem0 OSS + Qdrant · 56 comentarios #7847 docker-compose doesn't mount mem0/ as AGENTS.md claims · 4 comentarios #7888 Groq defaults still select retired models in the TypeScript and Python SDKs · 0 comentarios #7880 LLMReranker parses scientific-notation scores as 1.0, reversing search rankings · 4 comentarios #5958 perf(opensearch): N index()+refresh() round-trips per insert() — collapse to one bulk + one refresh · 3 comentarios #7879 bug(sdk-python): Memory.add() persists a vector despite SQLite history failure, creating duplicates on retry · 5 comentarios #4956 ADD-only extraction in v3 may surface stale/contradictory facts for time-sensitive attributes · 24 comentarios #7602 @mem0/opencode-plugin fails to load silently on OpenCode V2 (all published versions) · 7 comentarios #7871 docs(vector_stores): two constructors document arguments they do not accept · 1 comentarios #7810 `add()` accepts a non-string `created_at` in metadata, after which every read and delete for that user fails · 4 comentarios #7523 bug(llms): o4-mini is not classified as a reasoning model, so temperature/top_p/max_tokens are sent on every call · 6 comentarios #7202 Skill examples and LLM.md call search()/get_all() with top-level user_id, which raises ValueError · 7 comentarios