Category · Agent runtime
Memory layers for AI agents
Services that store what an agent learns about a user or a task and bring back the right parts later: facts, preferences, conversation summaries and knowledge graphs. Compared on recall quality, latency, what they keep and how you delete it.
Capability keys memory.store · memory.search · memory.graph · memory.user · memory.delete · All tools
letme.dev/memory.store picks the top-graded tool in this list and says how to call it direct; calling through letme comes later.
The same listing from the live API. Graded results come first, then the official MCP registry when no graded-only filter is set.
https://www.anchorterminal.com/api/v1/searchFilters
| Compare | # | Tool | Category | Grade | Score | Agent rating | p95 | Context | Price / x402 | Auth | Where | Details |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 112 | ZepZep · HTTP API | Memory | B | 69.6 | 4.0 (2) | n/a | n/a | $125 / mo | OAuth or key | Hosted | ||
|
Hosted context and memory service built on a temporal graph of facts, relationships and source episodes. Top strength Facts that new data invalidates keep the time they stopped being true Top weakness The terms of 17 August 2026 grant a perpetual, irrevocable licence to train models on customer data |
||||||||||||
| 188 | HonchoPlastic Labs · HTTP API | Memory | B | 64.2 | 3.0 (2) | n/a | n/a | $0.001 / call x402 | OAuth or key | Hosted | ||
|
Memory API that models each participant (a peer) in a conversation. Top strength Pay per call over x402 or MPP at agentcash.honcho.dev, with free read endpoints Top weakness No published rate limits, 429 guidance or SLA |
||||||||||||
| 201 | Supermemory API + MCPSupermemory · HTTP API | Memory | B | 63.6 | 3.5 (2) | n/a | n/a | $19 / mo | OAuth or key | Hosted | ||
|
Memory and context API that ingests text, URLs, PDFs, images and video, extracts memories into a graph per container tag (usually one per user) and returns them through search or a user profile endpoint. Top strength 8-tool hosted MCP with OAuth and per-space read or write permission Top weakness No legal entity named in the terms or privacy policy |
||||||||||||
| 301 | Mem0 Platform + MCPMem0 · HTTP API | Memory | C | 56.6 | 2.5 (2) | n/a | n/a | $19 / mo | API key | Hosted | ||
|
Hosted memory layer that extracts facts from conversations and returns the relevant ones for a user, agent or run on later turns. Top strength An agent can create its own Free account with Top weakness Free Plan data is used to train Mem0's models, per the privacy policy of 22 August 2026 |
||||||||||||
| 320 | CogneeCognee · Model platform | Memory | C | 54.6 | 3.0 (2) | n/a | n/a | $5 / mo | OAuth or key | Hosted + local | ||
|
Open-source memory engine that turns documents, conversations and synced sources into a knowledge graph plus a vector index and answers queries over both. Top strength Apache-2.0 library, REST server and MCP server, all self-hostable, with local models and no LLM key since 1.6.0 Top weakness No status page, rate limits, SLA or SOC 2 for Cloud, and Cloud runs in AWS us-east-1 with no EU region |
||||||||||||
| 322 | Memory (MCP reference server)MCP project (reference servers) · MCP server | Databases | C | 54.4 | 2.5 (2) | n/a | n/a | Free · OSS | None | Local | ||
|
Knowledge-graph persistent memory reference server (entities, relations, observations) stored as JSONL at MEMORY_FILE_PATH. Top strength Nine tools with typed schemas and accurate read, destructive and idempotent annotations Top weakness No pagination or limits. |
||||||||||||
| 333 | GraphitiZep · Agent framework | Memory | D | 53.3 | 2.5 (2) | n/a | n/a | Free · OSS | None | Library | ||
|
Open-source Python framework from Zep that builds a temporal knowledge graph from chat messages, text and JSON. Top strength Apache-2.0, with FalkorDB, Neo4j or Amazon Neptune as the store Top weakness Every add runs LLM extraction, so ingestion costs tokens and time |
||||||||||||
| 359 | HindsightVectorize · HTTP API | Memory | D | 50.4 | 3.0 (2) | n/a | n/a | $0.05 / call | OAuth or key | Hosted | ||
|
Memory engine for storing and retrieving information used by agents. Top strength Keys restricted to named banks, with expiry from an hour to a year and child-key revocation Top weakness Retain at $10 per million tokens is the priciest ingestion in this category |
||||||||||||
Nothing matches these filters. .
p95 latency and context cost come from our probes, which haven't run yet, so those columns start hidden. Grades run from AA to F, and agent-ready means BB or better. Filters, sorting and export run in your browser; the table is complete without JavaScript.
Indexed, not reviewed (63)
Listings sorted into this category from public catalogues (the official MCP registry, APIs.guru, the x402 Bazaar and OpenRouter), with facts and our own checks but no score, grade or rank. How the index works.
| Listing | Kind | What it does | Why it's here |
|---|---|---|---|
| 6DuckLearn MCP 6ducklearn.com | MCP server | Connect agents to 6DuckLearn memory, approvals, and runtime control. | vendor's own |
| Achriom achriom.com | MCP server | Media memory for AI agents and their humans: books, movies, music, shows, anime, podcasts, games. | vendor's own |
| adrkit decision memory adrkit.dev | MCP server | Deterministic, offline, read-only ADR decision memory for coding agents. No model or network calls. | vendor's own |
| AI Context Flow plurality.network | MCP server | Universal memory for AI agents and tools. Save, organize and search context anywhere. | vendor's own |
| Amber ambermem.com | MCP server | Long-term memory for AI assistants. Hybrid retrieval, query expansion, auto-topics. | vendor's own |
| Andish andish.ru | MCP server | Knowledge-graph memory for your team's documents. Answers cite the source, not a summary. | vendor's own |
| being-mcp-server ruddia.com | MCP server | Personality Runtime for AI agents — persistent memory, personality, and relationships via MCP | vendor's own |
| Bluet bluet.dev | MCP server | Product memory your agents read before they work: briefs, plan checks against decisions, debriefs. | vendor's own, widely used |
| Bourdon bourdon.ai | MCP server | Recognition-first cross-agent memory federation. One shared memory across all your agents. | vendor's own |
| br8n.io MCP server br8n.io | MCP server | Owned, portable working memory: a plain-files markdown brain any model can read and cite. | vendor's own |
| Brainfeather brainfeather.com | MCP server | Long-term memory for AI coding agents: durable project facts, recalled by every MCP client. | vendor's own |
| Cairn.ink Memory cairn.ink | MCP server | Cross-session memory for AI agents with a Source Receipt for every memory, over MCP. | vendor's own |
| Citadel citadeldb.dev | MCP server | Encrypted-first embedded database with vector search and agent memory, exposed as MCP tools | vendor's own |
| cluster-mcp clusteragent.dev | MCP server | Self-hostable index of AI financial agents — tokenized stocks + crypto, memory, LLM gateway. | vendor's own |
| Compose MCP smartmemory.ai | MCP server | Typed feature management for compose projects: roadmap, changelog, artifacts, journal. | vendor's own, widely used |
| ContextStream contextstream.io | MCP server | Persistent memory and cross-session learning for AI coding assistants (hosted remote MCP). | vendor's own |
| deerdawn deerdawn.com | MCP server | AI session memory: the brief your AI reads before every session so no session starts cold. | vendor's own |
| Docmancer docmancer.dev | MCP server | Local-only MCP server for source-attributed Markdown memory and documentation retrieval. | vendor's own |
| Engram engram.tools | MCP server | Memory for AI agent teams across tools, sessions, repositories, and teammates. | vendor's own |
| Engram getengram.app | MCP server | Persistent, verbatim, searchable memory for AI assistants — one memory across every MCP client. | vendor's own |
| engrava sovantica.ai | MCP server | MindQL queries, hybrid search, and a thought graph - read/write memory for AI agents. | vendor's own |
| gateway agent.wingmanprotocol.com | MCP server | Durable self for AI agents: one-call resume, memory, real browser, free chat + hire real humans. | vendor's own |
| gjalla gjalla.io | MCP server | Hosted self-curating shared memory that keeps your agents working like a high-performing team | vendor's own |
| Granoflow granoflow.com | MCP server | Connect MCP-capable AI agents to local Granoflow tasks, reviews, cards, imports, and work memory. | vendor's own |
| HyperMarrow qianshi.cool | MCP server | Local-first long-term memory for AI coding agents. Stored on your own machine. | vendor's own |
| in-memoria pi22by7.me | MCP server | Persistent codebase intelligence that gives AI assistants memory across sessions | vendor's own |
| kage kage-core.com | MCP server | Verified memory for coding agents: claims cited against code, stale withheld, savings receipts. | vendor's own |
| Kawa Code kawacode.ai | MCP server | Team-aware memory: intent, decisions, real-time conflicts for AI coding assistants. | vendor's own |
| Koragraph koragraph.in | MCP server | Local multi-repo code graph over MCP. Blast radius, call graphs, and memory anchored to code. | vendor's own |
| Kortexio Memory kortexio.io | MCP server | Hosted ambient memory for AI agents. Retrieve context and save notes via MCP tools with OAuth. | vendor's own |
| Lex smartergpt.dev | MCP server | Episodic memory, Frames, Policy Neighborhoods, and policy enforcement for AI agents. | vendor's own |
| Lithtrix — Identity, Memory & Trust for AI Agents lithtrix.ai | MCP server | Identity, memory, trust, and swarm primitives for AI agents. MIRC base always free. | vendor's own |
| Living Memory viibe.to | MCP server | An MCP memory server. One memory your agents share — across models, devices and apps. | vendor's own |
| Lodekeep lodekeep.com | MCP server | Memory MCP server for Claude Code. Correct a fact and the old version stops being recalled. | vendor's own |
| logic-server nocturnus.ai | MCP server | Agent reasoning, memory, and token-optimized context for AI applications. | vendor's own |
| Magnemo magnemo.ai | MCP server | Governed memory for AI agents. Memory with receipts. Local, plain files, zero network calls. | vendor's own |
| Marrow marrow.id | MCP server | Evidence-grounded, graph-connected, correctable memory for agents. | vendor's own |
| MemCell memcell.ai | MCP server | Living memory for AI coding agents: recall before acting, report outcomes so confidence is earned. | vendor's own |
| memento runmemento.com | MCP server | Local-first, LLM-agnostic memory layer for AI assistants. | vendor's own |
| memgit memgit.dev | MCP server | Git for AI memory — version-controlled, searchable context that persists across sessions | vendor's own |
| MemoFS MCP Server memofs.dev | MCP server | Persistent memory and virtual filesystem MCP server for AI agents. | vendor's own |
| memory cortex-mem.com | MCP server | Shared, persistent memory for AI agents — works with Claude, Kiro, Cursor, or any MCP client. | vendor's own |
| memory-mcp letta.com | MCP server | MCP server for AI memory management using Letta - Standard MCP format | vendor's own |
| MemoryLayer memorylayer.in | MCP server | Local-first persistent memory for AI agents, with reasoning and formal verification. | vendor's own |
| memxus memxus.com | MCP server | Persistent memory layer that saves and recalls your project context and preferences. | vendor's own |
| mnemodb mnemodb.dev | MCP server | Persistent agent memory as plain Markdown in your repo - readable, diffable, with provenance. | vendor's own |
| Neonia: The Cloud Backend for AI Agents neonia.io | MCP server | Multiple MCP tools, persistent graph memory, token-saving data pointers, and more. | vendor's own |
| Operations Pulse openlyuseful.org | MCP server | Local-first operational pulses, durable ticket memory, and governed connection discovery. | vendor's own |
| Pastepile pastepile.com | MCP server | Durable, versioned memory for AI assistants. Remote memory needs a Pastepile subscription. | vendor's own |
| RCLL rcll.ai | MCP server | Self-hosted shared memory for a team of AI agents. Rooms, L0-L3 depth, no LLM on the read path. | vendor's own |
| ReasonGraph Cloud memory memory.primaxiom.ai | MCP server | Graph memory for AI agents: entities, cause-effect links, cross-session recall, time travel. | vendor's own |
| recall clauderecall.com | MCP server | Persistent memory for Claude Code and Codex. Search and re-inject context from any session. | vendor's own |
| Reverie knowall.ai | MCP server | Graph memory that dreams: Neo4j knowledge-graph memory for AI agents with hybrid search | vendor's own |
| sdk mnemopay.com | MCP server | Memory + wallet for AI agents. Real payment rails, Agent FICO 300-850, Merkle audit, identity. | vendor's own |
| stdio-proxy mcp.undisk.app | MCP server | Undisk MCP: safe file memory for AI agents. | vendor's own |
| Synap Memory maximem.ai | MCP server | Persistent memory for AI agents — log and recall conversation context over MCP. | vendor's own |
| Synchronex synchronex.ai | MCP server | Connect any AI agent to your Synchronex workspace — portable memory, decisions, and workers. | vendor's own |
| Trail usetrail.dev | MCP server | Issue tracker and durable project memory for your coding agent, scoped to the folder you open. | vendor's own |
| Tugra tugra-ai.com | MCP server | Provenance-aware memory format for AI agents: every claim carries its source, age and boundary. | vendor's own |
| wontopos.com MCP server wontopos.com | MCP server | Wontopos (WOS) long-term memory (beta) - one memory across Claude Code, Cursor, and your agents. | vendor's own |
The first 60 of 63, most used first. All 63 are in this page's JSON twin and, with their facts, in /api/v1/indexed.json (category).
How we test this category
One user's history fed in over several sessions, then questions that need facts from early on, a changed preference and a deleted fact. We check what comes back, how fast, and whether the deletion sticks. This test hasn't run yet, so Task success is pending and the grades here come from the categories assessed from public evidence.
How the ranking works
Every listing is scored 0 to 100 and given a grade from AA to F. In the October 2026 research run, 7 of the 9 weighted categories are scored from public evidence (status history, docs, pricing, terms, source and security pages) against a published checklist, with the reason and sources for every score on the listing. Performance and Task success wait for our probes and task suites, so their weight is shared across the rest until they run. Negative events deduct up to 15 points. Read the methodology.
For companies
Do agents find, use and choose your tools?
An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.