Catalog Systems Agent frameworks (no first-party memory layer) Model Context Protocol (MCP spec)

Model Context Protocol (MCP spec)

Anthropic-introduced open protocol (Nov 2024) for connecting LLMs to data sources and tools. Now standard across Anthropic, OpenAI, Microsoft, Cursor, Continue, IDEs, etc.

At a glance

Type: Open protocol for LLM-tool I/O
Tier: T1
Section: Agent frameworks (no first-party memory layer)
Created: 2024
Latest release: searched not found
License: MIT (spec)
GitHub: searched not found
Pricing: searched not found
Funding: searched not found

Taxonomy

storage: n/a
retrieval: n/a
persistence: n/a
update: n/a
unit: n/a
governance: n/a
conflict: n/a

When to use

Optimised for: searched not found

Anti-fit: searched not found

Pros & cons

Pros

searched not found

Cons

searched not found

Claims & capabilities

Adopted by OpenAI, Microsoft, Cursor, Continue, Windsurf within ~6 months of release.

Technical surface

API surface: searched not found
Backend storage: not applicable — not a memory product
Deployment: searched not found
Embedding model: not applicable — not a memory product
Multi-tenancy: searched not found
MCP: searched not found
A2A: searched not found
OpenTelemetry: searched not found

Compare Model Context Protocol (MCP spec) with…

Similar systems

Other agent frameworks (no first-party memory layer) in the catalog, ranked by inbound references.

LangChain (framework) T1
Dominant Python/JS framework for LLM apps — tool calling, agents, retrievers, runnables. Memory is one component (LangMem cross-listed); the framework itself is the product.
LangGraph (orchestration) T1
LangChain's stateful graph runtime for agents — explicit state, persistence, time-travel debugging, human-in-the-loop. Memory is one persistence target (LangGraph Persistence cross-listed).
LlamaIndex (framework) T1
Data framework for LLM apps — connectors, indices, query engines, agents, workflows. Memory is one of many subsystems (LlamaIndex Memory cross-listed).
OpenAI Swarm (deprecated) T4
OpenAI Swarm — experimental educational multi-agent framework released Oct 2024; deprecated in favour of the OpenAI Agents SDK in Mar 2025.
Strands Agents (AWS) T3
AWS-released open-source agent framework — released March 2025; model-driven, simple Agent abstraction. Predecessor of AWS Agent SDK / Bedrock AgentCore.
Agency Swarm T3
Agent framework built on OpenAI Assistants API — agencies (roles + comm flow); replaced after Assistants v2 changes.

Related systems

Referenced by (34)

Alita depends on at runtime — imal self-evolution via MCP generation. Headline: 75.15% pass@1 and 87.27% pass@3 on GAIA validation; 74.00%
Alita-G depends on at runtime — st to domain expert via MCP synthesis, abstraction, and curation into MCP Box. Headline: 83.03% p
Cipher (byterover) depends on at runtime — aude Code, Roo Code via MCP. Maintains a coding-context memory across sessions and across IDEs.
claude-brain depends on at runtime — adjacent-infrastructure cell: requires Claude Code or MCP client; BYO storage
Claude Code (Anthropic) depends on at runtime — adjacent-infrastructure cell: Anthropic API; MCP server ecosystem
Claude Code auto-memory depends on at runtime — adjacent-infrastructure cell: requires Claude Code or MCP client; BYO storage
Claude-CursorMemoryMCP depends on at runtime — adjacent-infrastructure cell: requires Claude Code or MCP client; BYO storage
claude-mem depends on at runtime — adjacent-infrastructure cell: requires Claude Code or MCP client; BYO storage
claude-supermemory depends on at runtime — adjacent-infrastructure cell: requires Claude Code or MCP client; BYO storage
Cline (harness) depends on at runtime — adjacent-infrastructure cell: OpenAI-compatible API; MCP
codebase-memory-mcp (DeusData) depends on at runtime — adjacent-infrastructure cell: self-contained binary + SQLite WAL; MCP client (Claude Code/Codex CLI/etc.); no LLM lock-in
Cognee MCP depends on at runtime — adjacent-infrastructure cell: requires Claude Code or MCP client; BYO storage
Continue.dev (harness) depends on at runtime — adjacent-infrastructure cell: any OpenAI-compatible API; MCP
Continue.dev Memory MCP depends on at runtime — Memory is injected via MCP rather than native to the Continue core. Scope depends on MCP server deploymen
Coveo depends on at runtime — for AWS (Dec 2025) via hosted MCP server grounding Amazon Bedrock agents in enterprise knowledge. Passage retr
engram depends on at runtime — adjacent-infrastructure cell: requires Claude Code or MCP client; BYO storage
Gemini Enterprise Agent Platform Memory Bank (rebrand of Vertex AI Memory Bank) depends on at runtime — adjacent-infrastructure cell: requires GCP (Vertex AI / Gemini Enterprise); on-prem option via open-source Gemma 4; managed MCP servers; Pub/Sub + Cloud Trace + IAM
Goose (Block) depends on at runtime — adjacent-infrastructure cell: LiteLLM for multi-model routing; MCP for tool integrations; Block-internal extensions (square/cash app workflows)
Graphiti MCP Server (Zep) depends on at runtime — adjacent-infrastructure cell: requires Claude Code or MCP client; BYO storage
Hindsight MCP depends on at runtime — adjacent-infrastructure cell: requires Claude Code or MCP client; BYO storage
Kuzu depends on at runtime — Sub-3ms recall via MCP.
local-memory-mcp depends on at runtime — adjacent-infrastructure cell: requires Claude Code or MCP client; BYO storage
mcp-knowledge-graph depends on at runtime — adjacent-infrastructure cell: requires Claude Code or MCP client; BYO storage
mcp-memory (Puliczek) depends on at runtime — adjacent-infrastructure cell: requires Claude Code or MCP client; BYO storage
mcp-memory-service (doobidoo) depends on at runtime — adjacent-infrastructure cell: requires Claude Code or MCP client; BYO storage
Mem0 MCP (official) depends on at runtime — hat brings the Mem0 layer into any MCP-compatible agent — including Claude Code — without rewriting code.
Memobase depends on at runtime — r Claude / Cursor / MCP-compatible agents.
Official MCP Memory server depends on at runtime — adjacent-infrastructure cell: requires Claude Code or MCP client; BYO storage
OpenAI Codex CLI depends on at runtime — adjacent-infrastructure cell: OpenAI API (Codex / o-series); MCP for tool extensibility; runs locally with TLS-only network egress
Opencode (SST) depends on at runtime — cript, multi-model, MCP-native. Released 2025; explicitly positioned as a permissively-lic
OpenMemory MCP depends on at runtime — adjacent-infrastructure cell: requires Claude Code or MCP client; BYO storage
Pieces for Developers (LTM) depends on at runtime — on-device, exposes via MCP for any IDE/agent.
RL Developer Memory depends on at runtime — adjacent-infrastructure cell: Model Context Protocol (MCP)-native architecture; logged contextual-bandit policy gates; static (11/11) and dynamic (10/10) integration test harness
Superpowers episodic-memory plugin depends on at runtime — s search/show tools via MCP. Distributed via obra/superpowers-marketplace ; install with /plugin install episodic-memory@superpowers-