"OCR agent" matching MCP servers:

Enables hybrid search over policies using Reciprocal Rank Fusion and provides grounded, context-aware answers via a LangGraph agent with COSTAR prompting.

A multi-agent Retrieval-Augmented Generation system exposed as an MCP server. Ask a question and a LangGraph pipeline plans the retrieval, pulls evidence from a pgvector knowledge base, optionally augments it with live web research, drafts a cited answer, and then self-critiques it for grounding — revising until the answer is supported by the sources.

Embedded, local-first agent memory: facts extracted into a per-namespace SQLite file (vec0 + FTS5) with hybrid retrieval and point-in-time (time-travel) queries. ADD-only history over stdio — no server process, no cloud dependency.

A personal memory MCP server that ingests AI agent conversation logs from multiple platforms into a searchable PostgreSQL+pgvector database, enabling cross-session recall of past reasoning and decisions.

A local-first knowledge base for LLM coding agents that indexes repository documentation, concept ontology, and build targets into Qdrant and exposes retrieval as MCP tools (search, get, list sources, reindex).

Enables ingestion and semantic search over text documents using PostgreSQL + pgvector and OpenAI-compatible embeddings, allowing any LLM agent to retrieve relevant chunks for grounded answers.

Enables AI agents to maintain long-term, cross-session memory by extracting facts, reconciling state conflicts, and retrieving relevant memories via vector search.

Gives AI agents private, persistent, semantically searchable long-term memory with RAG, governed multi-agent scopes, and real-time memory-change subscriptions.

Local-first agent-memory MCP server with a why() tool: recall a fact together with its connected subgraph (multi-hop), so linked memories surface even when they share no words with the query. remember/recall/relate/forget/why over one fused vector + graph + columnar engine a single offline Rust binary.

Integrates R2R (Retrieval-Augmented Generation) with Claude Desktop, enabling semantic search across knowledge bases and RAG-based question answering with support for vector, graph, web, and document search.

Local memory for AI agents over plain Markdown files: hybrid vector + BM25 recall (RRF), one isolated memory per agent directory, fully offline with no LLM in the loop. Tools: cogvault_recall and cogvault_record.

Self-hosted MCP-native agent memory server. Gives AI agents persistent, decay-weighted memory via 83 MCP tools — no cloud, full control. RocksDB+HNSW backend. Works with Claude Code, Cursor, and any MCP-compatible agent.

A TypeScript-based MCP server that provides project-specific knowledge graph memory for LLM agents to store and retrieve entities, relations, and observations. It features disk-persistent storage and supports cross-project knowledge sharing to enhance agent long-term memory.

VecFS is a lightweight, local-first vector storage specification and implementation for AI agent long-term memory. It provides an MCP server enabling agents to search, memorize, and manage context.

Provides persistent memory management for AI agents with 14 tools, including semantic search, context synthesis, auto-context middleware, and write-back actions to external services like Slack.

Enables semantic search and contextual conversations with your Calibre ebook library using vector-based RAG technology. Supports project-based organization, multi-format book processing, and OCR capabilities for enhanced content extraction and retrieval.

It connects an agent to a PostgreSQL database so it can inspect schemas, run validated reads and writes, resolve fuzzy names via trigram search, and perform or store pgvector embedding searches. Vector and hybrid search operate through an OpenAI-compatible embeddings endpoint, letting agents query by text without handling embedding arrays.

Enables AI agents to read, write, organize, and link persistent long-term memory across model providers using embeddings and pgvector retrieval.

Tillpad provides bounded storage and search for agent jobs: bounded KVP storage, RAG search, and wipe receipts so long-running agents keep durable, size-capped context.

A local-first AI backend that combines Model Context Protocol, Retrieval-Augmented Generation, and an agent runtime to answer questions from a private knowledge base.