Skip to main content
Glama
fish827-08

kb-memory

by fish827-08
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Python
Hybrid

Related Servers

Alternatives to kb-memory

No user-submitted related servers found.

    Related Servers

    • mnemostack

      udjin-labs
      A
      license
      A
      quality
      A
      maintenance
      Durable hybrid memory for AI agents. Combines vector search, BM25, temporal retrieval, and optional Memgraph knowledge graph via reciprocal rank fusion. 6 MCP tools: health, search, answer, feedback, graph_query, graph_add_triple. Self-hosted with Qdrant backend.
      7
      51 PyPI
      7
      Apache 2.0

    • llm-memory-tool

      menoxz
      A
      license
      Not graded
      quality
      C
      maintenance
      MCP server that enables persistent, hybrid, local memory for LLM agents, with vector + BM25 search, knowledge graph, and policy-driven retention, providing token-budgeted context injection for AI assistants.
      MIT

    • HCC

      icemaple77
      A
      license
      B
      quality
      B
      maintenance
      Local-first, cross-agent memory layer for Hermes, OpenClaw, and Claude Code (MCP) behind a single REST API. Hybrid BM25+pgvector+RRF retrieval with optional Qwen3 rerank, a three-stage "dream" consolidation cycle (Light → REM → Deep) that clusters and de-dupes daily memories into durable knowledge with a narrative journal written to Obsidian, and a persistent 6-dimension emotion state machine that
      8
      1
      MIT

    • Strata

      avinash-1707
      A
      license
      Not graded
      quality
      B
      maintenance
      Local-first memory layer for AI agents. Compresses context with a local LLM and serves hybrid search over MCP.
      521 npm
      MIT

    • RecallNest

      AliceLJY
      A
      license
      A
      quality
      A
      maintenance
      Persistent memory MCP server for AI coding agents (Claude Code, Codex, Gemini CLI). Hybrid retrieval (vector + BM25), cross-encoder reranking, knowledge graph, session checkpoint/resume, and multi-scope isolation. Local-first with LanceDB.
      30
      205 npm
      15
      MIT
    Deploy Server
    One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

    TDQS

    A4.2/5.0

    Scored across 8 tools

    Disambiguation5/5

    Each tool targets a distinct resource and action: CRUD operations for memories are clearly separated from hybrid retrieval, and the two ingestion tools are differentiated by source type (local file vs URL). search_memory and ask_kb are distinguishable by their outputs—raw hits versus a generated answer with sources.

    Naming Consistency5/5

    All tools follow a consistent verb_noun snake_case pattern (write_memory, read_memory, update_memory, delete_memory, search_memory, add_document, add_webpage, ask_kb). The only slight deviation is ask_kb's abbreviated noun, but it still fits the verb_noun structure and is readable.

    Tool Count5/5

    8 tools is well-scoped for a memory/knowledge server: full CRUD for memories, two ingestion tools for shared knowledge, hybrid search, and RAG Q&A each earn their place. The count is neither sparse nor bloated.

    Completeness4/5

    The memory lifecycle is fully covered (write/read/update/delete), and knowledge ingestion plus retrieval and Q&A are present. Minor gaps exist—there is no list/delete operation for shared documents or webpages—but agents can work around these since shared knowledge is append-only.

    Maintenance

    ActivityMaintained
    ResponsivenessNo issues

    Resources