1. Conduid
  2. RAG
  3. Hebbrix MCP
MCP server · RAG

Hebbrix MCP

Hebbrix MCP server — long-term memory and knowledge graph for any MCP-compatible agent (Claude Desktop, Cline, Cursor, Continue)

34Low

Scored a month ago · breakdown

About Hebbrix MCP

Hebbrix MCP is an MCP server in the RAG category: hebbrix MCP server — long-term memory and knowledge graph for any MCP-compatible agent (Claude Desktop, Cline, Cursor, Continue).

Install

From source
git clone https://github.com/Hebbrix/hebbrix-mcp

This server has no ConduID identity, so agent calls to it are not receipted. Pin the version you install and review the source before granting it credentials.

Tools (29)

hebbrix_account_statusTier, usage, limits, and expiry.

Parameters not published.

hebbrix_ask**One-call GraphRAG.** Ask a natural-language question; it searches memory, synthesizes an answer with an LLM, and cites the memory ids it used. If synthesis abstains while authoritative search has grounded evidence, it returns that evidence as `synthesis_status: "retrieval_only"` instead of silently losing recall or pretending synthesis succeeded.

Parameters not published.

hebbrix_choose_actionSafely choose among repeatable strategies and create

Parameters not published.

hebbrix_confidenceHow confident should the agent be before acting? Grounded in memory + past outcomes.

Parameters not published.

hebbrix_contradictionsSurface facts that conflict with each other.

Parameters not published.

hebbrix_create_procedureStore a scoped condition/action procedure.

Parameters not published.

hebbrix_delete_procedureIdempotently delete a procedure and its executions. The API returns the same 204 for deleted, absent, and foreign-tenant IDs so the tool cannot reveal another tenant's identifier.

Parameters not published.

hebbrix_entity_timelineWhat was true about an entity, and when.

Parameters not published.

hebbrix_execute_procedureExecute a procedure and record the execution.

Parameters not published.

hebbrix_exportExport a whole collection (memories + graph entities + profile) as JSON or Markdown, in one call.

Parameters not published.

Show all 29 tools
hebbrix_extraction_statusPoll a smart-ingestion job until its created/updated memories or terminal error are available.

Parameters not published.

hebbrix_forgetDelete a memory by id.

Parameters not published.

hebbrix_getFetch one memory by id, with metadata.

Parameters not published.

hebbrix_graph_queryTraverse relationships out from a named entity; pass a `timestamp` for point-in-time truth. Results are trimmed (from/to/type/valid_from), not raw backend payloads. (Free-text questions: use `hebbrix_ask`.)

Parameters not published.

hebbrix_graph_statusCheck one memory's asynchronous graph readiness, optionally polling for up to 30 seconds. Readiness comes from durable, source-revision-bound delivery acknowledgements—not elapsed time, search readiness, or an empty graph query. It distinguishes successful enrichment with no related facts from a still-processing, failed, or temporarily unreadable graph.

Parameters not published.

hebbrix_historySee how a memory changed over time.

Parameters not published.

hebbrix_importThe inverse of export: import a list of facts, an export JSON, or notes/markdown into a collection (restore a backup, migrate, or seed from `CLAUDE.md`).

Parameters not published.

hebbrix_learning_insightsInspect posterior probabilities, credible

Parameters not published.

hebbrix_listList recent memories.

Parameters not published.

hebbrix_list_collectionsList the memory spaces this key can use.

Parameters not published.

hebbrix_log_decisionRecord a decision and its outcome; feeds future confidence. Right after a `hebbrix_confidence` check you can log just the `outcome` — the description auto-fills from what you asked.

Parameters not published.

hebbrix_mark_usedReinforce a memory you actually used (`helpful=True` strengthens it, `False` weakens it) so recall improves over time.

Parameters not published.

hebbrix_rememberStore a fact, decision, or preference.

Parameters not published.

hebbrix_remember_manyStore **many** facts in one call (one round-trip, one rate-limit hit). Pass `facts` (list of strings). Falls back to sequential writes on free/agent tiers.

Parameters not published.

hebbrix_report_outcomeClose that decision loop later with `success`, a

Parameters not published.

hebbrix_searchSemantic search (hybrid vector + BM25 + graph retrieval).

Parameters not published.

hebbrix_search_entitiesList known entities (people, orgs, tools, places).

Parameters not published.

hebbrix_updateCorrect a memory **in place** (old versions are kept).

Parameters not published.

hebbrix_update_procedureUpdate mutable fields; ownership scope is immutable.

Parameters not published.

Ask AI

✦

Ask AI about Hebbrix MCP

Powered by Claude · Grounded in docs

I know everything about Hebbrix MCP. Ask me about installation, configuration, usage, or troubleshooting.

Security checks

6 checks have not been run yet.

Questions

About Hebbrix MCP

How do I install Hebbrix MCP?

No package has been published for this server, so install it from source: clone the repository and follow its README.

Is Hebbrix MCP safe to use with an AI agent?

Its trust score is 34 out of 100 (low). Conduid hasn't run static security checks on this repository yet, so review the source yourself before granting it credentials. It has no ConduID identity yet, so agent calls to it are not receipted.

Is Hebbrix MCP still maintained?

Conduid hasn't recorded a commit date for this repository yet. Check the repository directly for recent activity.