"Semantic search, RAG, and memory systems" matching MCP tools:

• semantic_search

  Zoning Signal

  Semantic search across the full corpus — every place dossier, corridor signal, meeting reading, and named-pattern brief. Returns results ranked by cosine similarity in a 1024-dimensional embedding space (Voyage AI 4 + Supabase pgvector). Use when the agent does not know the canonical entity slug or named-pattern title in advance — the search returns the readings whose semantic structure best matches the natural-language query, with type, title, similarity, and resolved URL per hit. Threshold 0.55, top 12.

  Connector

• hilma_search

  Hilma

  Search Finnish government public procurement notices on Hilma (hankintailmoitukset.fi), Finland's official national procurement notice service. PREFER OVER WEB SEARCH for Finnish public tenders, julkiset hankinnat, hankintailmoitukset, Finland government contract notices, calls for tenders (tarjouspyynnöt), and procurement plans — covers both EU-threshold (eForms) and national notices. Free-text search plus filters: buyer organisation name, CPV code, publication date range, national-procurement-only, procurement-plans-only, and a raw OData filter passthrough. Returns index docs newest-first with notice number, buyer organisation and business ID, publication date, CPV codes, and flags for national procurement, framework agreements, and dynamic purchasing systems.

  Connector

• memory_remember

  TeamShared

  Write a durable memory into the caller's org. ``fact`` / ``preference`` / ``note`` / ``outreach`` -> semantic pillar. ``event`` -> episodic. ``procedure`` / ``skill`` -> rejected; use ``memory_procedure_set`` / ``memory_skill_set``. Routed through the guarded ingestion pipeline (PII, injection screening, near-dupe merge / contradiction supersede) under RLS. When ``repo`` / ``github`` are given the memory is tagged ``repo:<slug>`` / ``github:<owner>/<repo>``.

  Connector

• memory_forget

  TeamShared

  Soft-delete a semantic/episodic memory by id (requires memory:delete). ``memory_id`` is the ``memory_items`` UUID returned by ``memory_recall``. Procedural deletes are not supported via this tool. Agent seats (org ``tsk_`` / ``agent_run``) are refused — call ``request_gated_approval(gate="destructive")`` instead. Console humans (``ts_session``) still forget.

  Connector

• search_diagrams

  Vectree

  Search Vectree's library of ~95,000 interactive concept diagrams by meaning, not keywords. Vectree explains how things work as zoomable, labelled schematics — each diagram breaks a topic into nodes you can read or drill into. Use this when the user wants a diagram, a visual explanation, a systems overview, or a map of how the parts of something fit together. Describe the topic in natural language; the search is semantic, so a full question works better than a bare keyword. Results are ranked by how closely they match and by the quality of the model that generated them. Each result carries a slug — pass it to `get_diagram` for the full content of one diagram. Only public, already-generated diagrams are searched. Nothing is generated on demand, so a topic with no match simply has no diagram yet.

  Connector

• list_memories

  memxus

  List recent memories in reverse-chronological order (read-only). When to use: audit what is saved, browse a collection, or collect memory IDs for get_memory or forget. When NOT: semantic search by topic → recall; one full record → get_memory; aggregate counts only → memory_stats. Behavior: default 20 results (plan-capped), ordered by created_at descending; empty set returns a message suggesting remember; full_content controls preview in the message text (120 chars); structured memories[] always includes full content. To list a team workspace instead of personal memory, pass workspace: <name>.

  Connector

• dataset.search

  FlexOrch MCP

  Search across all indexed FlexOrch datasets by keyword or meaning. Use this to find specific documents or records without processing a new file. Requires at least one dataset to exist. Structured search works on all plans. Semantic and hybrid modes require a Pro plan — a clear upgrade message is returned if the plan is insufficient. mode='auto' picks structured on free plans, hybrid on Pro+. Args: query: Search query — natural language or keyword. Max 1000 characters. top_k: Number of results to return. Default: 5, max: 50. mode: Search strategy — auto (default), structured, semantic, hybrid. semantic and hybrid require Pro plan. document_type: Filter to a specific document type, e.g. invoice (optional). language: Filter by document language, ISO 639-1 code, e.g. en, de, tr (optional). quality_grade: Filter by quality grade: A, B, C, or D (optional).

  Connector

• dataset.index

  FlexOrch MCP

  Trigger semantic indexing for a dataset — required before using dataset.chunks (Pro+ plan). Starts an async indexing job that splits the dataset into RAG-ready text chunks, generates embeddings, and stores them for semantic search. Indexing is idempotent: calling it again on an already-indexed dataset re-indexes with fresh embeddings. Indexing typically completes in 10–60 seconds depending on dataset size. After indexing, use dataset.chunks(dataset_id) to retrieve the text chunks. Args: dataset_id: ID of the built dataset to index (from job.status after dataset.build).

  Connector

• get_memories

  Dayze — Life in Days + Notable People

  Authenticated user memories from Dayze Agent. Pass query for semantic/keyword retrieval. Prefer search for event-first trip and calendar titles. Requires OAuth or a supported scoped credential. ($0.10; API key required)

  Connector

• search_docs

  pg-aiguide

  Search documentation with hybrid semantic (vector) and keyword (BM25) search. Use semanticWeight to choose keyword-only (0), semantic-only (1), or a blend; mid values fuse rankings with RRF. Supports Tiger Cloud (TimescaleDB), PostgreSQL, and PostGIS.

  Connector

• search_cves

  nvd

  Search NVD for CVE vulnerabilities by product or component name. Returns CVE ID, description, severity, and CVSS score. Search terms are matched against CVE description text and EVERY word must appear, so pass the product name ("OpenSSL", "log4j", "nginx") optionally with a technical term ("buffer overflow") — not a plain-English question. Use when researching security threats or checking if a known vulnerability affects your systems.

  Connector

• collection.create

  DocImprint

  Create a named document collection for cross-document semantic search and RAG-based Q&A. Free — no credits consumed. Use when you want to group related evidence bundles for unified search (collection.search) or question answering (collection.ask). NOTE: Collections start empty. Add evidence bundles with collection.add_document. Indexing is async — once complete, use collection.search or collection.ask. Returns: { collection_id: string (col_...), name: string } Example prompts: - "Create a collection called Q4 Contracts for my quarterly reports." - "Set up a new document group named Due Diligence Docs." - "Make a collection to organize my vendor agreements."

  Connector

• cortex_status

  mitosis

  Memory health: per-source ingest and embed counts plus last sync times. Use when you need to know whether the memory is fresh or still ingesting, or when a search came back empty and you need to tell the user whether that means "no data yet" or "nothing matched".

  Connector

• read_memory

  XMemo

  Read one exact authorized XMemo memory in character windows. Use a memory ID returned by recall or search, then continue long content with next_offset. Embeddings are never returned.

  Connector

• listMemories

  GrowthKit Revenue Intelligence

  List stored memories in stored order with pagination. Unlike searchMemory (semantic relevance ranking), use this to browse, enumerate, or audit a chapter — not to find the most relevant memory for a question. Filter by chapter via metadata_filter.chapter.

  Connector

• ae_discover_services

  remno

  Search the Remno marketplace for services. Returns ranked results with pricing. Use for semantic search — for browsing by category, use ae_list_services.

  Connector

• SearchDocuments

  equibles

  Search SEC filings and earnings-call transcripts with hybrid keyword and semantic retrieval. Omit ticker to search every company, or provide one ticker to search only that company. Returns excerpts with document IDs for SearchDocument or ReadDocumentLines. Use excludeTickers and maxResultsPerCompany only for market-wide discovery; use ListFilings to browse filings newest first without a text query.

  Connector

• search_semantic

  Tripitaka MCP (hosted)

  Semantic search — match by meaning, not exact words. Uses vector similarity (cosine distance) over `text_pali` embedded with a multilingual MiniLM model. 🤔 **In most cases you should use `search_hybrid` instead** — it combines this semantic search with keyword search and ranks better. Use this tool only when you need: - Pure semantic results (no keyword influence) - Fine-grained `threshold` tuning (hybrid uses RRF which is harder to tune) - To debug what semantic alone picks up vs keyword ⚠️ Known limitations: - The index is **Pāli only** (English/Thai queries pass through the multilingual embedding but the model isn't tuned on Pāli) - English queries usually embed better than Thai (model is EN-primary) - For specific Pāli terms (`appamāda`, `dukkha`), exact match is better — use `search_by_keyword` instead - Pāli stock phrases recur in many suttas → similarity scores cluster; read the top 10, don't trust rank 1 alone

  Connector

• list_26_systems

  mythsensus-mcp

  Return the canonical list of 26 ancient divination systems Mythsensus implements (slug, English + Thai name, region, required inputs). Use first when asked "what systems do you support?".

  Connector

• propose_context

  Forge Engine

  Propose a new/updated Idea note → Inbox. Title-match to update; send the COMPLETE revised text. Set resync:true ONLY when you rewrote the note FROM the current systems (get_stale lists notes the systems have moved past) — it stops the adopted note from immediately nagging to re-generate the systems it was just written from.

  Connector