# Best memory layers for AI agents

> Amazon Bedrock AgentCore Memory (A), Zep (B) and Honcho (B) lead the 10 ranked memory layers. Picks by need, strengths, weaknesses and prices from the Anchor benchmark.

- Canonical: https://www.anchorterminal.com/best/agent-memory/
- Markdown: https://www.anchorterminal.com/best/agent-memory/index.md (~5,300 tokens)
- Slim: https://www.anchorterminal.com/best/agent-memory/index.min.md (~1,330 tokens, same facts, less prose, for token-sensitive contexts)
- JSON: https://www.anchorterminal.com/best/agent-memory/index.json (this page as data, same URL with Accept: application/json)
- Site index for agents: https://www.anchorterminal.com/llms.txt (full text: https://www.anchorterminal.com/llms-full.txt)
- API: https://www.anchorterminal.com/api/v1/index.json
- Updated: 2026-10-08

All 10 ranked memory layers on the Anchor benchmark, with a pick for each need and where each one falls short. Scores come from public evidence, re-checked as vendors change.

- Ranked: 10 · agent-ready (BB or better): 1 · accept x402: 1 · hosted endpoints: 7
- Full ranked table: https://www.anchorterminal.com/categories/agent-memory.md
- Head-to-head comparisons: https://www.anchorterminal.com/compare/agent-memory/index.md (55)
- Methodology: https://www.anchorterminal.com/benchmark/index.md

## The shortlist

| # | Tool | Grade | Score | Best for | Price | Where |
| --- | --- | --- | --- | --- | --- | --- |
| 1 | [Amazon Bedrock AgentCore Memory](https://www.anchorterminal.com/tools/agentcore-memory.md) | A | 79.4 | Teams already on AWS that want per-user memory under IAM, KMS and Regional controls, with extraction run for them. | $1 / GB | hosted and local |
| 2 | [Zep](https://www.anchorterminal.com/tools/zep.md) | B | 69.3 | Agents that must track how facts about a person or account change over time, inside a team that needs access policies and audit logs. | $125 / mo | hosted |
| 3 | [Honcho](https://www.anchorterminal.com/tools/honcho.md) | B | 64.1 | Products that model the people in a conversation and want to ask questions about them, and for agents that must pay their own way. | $0.001 / call | hosted |
| 4 | [Supermemory API + MCP](https://www.anchorterminal.com/tools/supermemory.md) | B | 63.6 | Apps that need user memory and document search over the same data, and for multi-tenant products that want a key per user. | $19 / mo | hosted |
| 5 | [Mem0 Platform + MCP](https://www.anchorterminal.com/tools/mem0.md) | C | 56.1 | Chat products that want per-user facts back with one search call and little setup. | $19 / mo | hosted |
| 6 | [Cognee](https://www.anchorterminal.com/tools/cognee.md) | C | 54.6 | Agents whose memory has to include documents, wikis and chat tools as well as conversation, and for teams happy to self-host. | $5 / mo | hosted and local |
| 7 | [Memory (MCP reference server)](https://www.anchorterminal.com/tools/memory-reference-server.md) | C | 54.2 | A single local agent that wants a small, inspectable store of facts about people and projects. | Free · OSS | local |
| 8 | [Graphiti](https://www.anchorterminal.com/tools/graphiti.md) | D | 53.3 | Teams that want Zep's temporal graph model on their own infrastructure and can run a graph database. | Free · OSS | library |
| 9 | [Hindsight](https://www.anchorterminal.com/tools/hindsight.md) | D | 50.2 | Agents that should form opinions and summaries from what they stored, such as long-running assistants or copilots that reflect on past sessions. | $0.05 / call | hosted |
| 10 | [LangMem](https://www.anchorterminal.com/tools/langmem.md) | D | 47.9 | Teams already on LangGraph that want memory tools and background extraction over a store they run. | Free · OSS | local |

## Picks by need

- Highest score overall: [Amazon Bedrock AgentCore Memory](https://www.anchorterminal.com/tools/agentcore-memory.md), A, 79.4/100 on the benchmark. Also [Zep](https://www.anchorterminal.com/tools/zep.md), B, 69.3/100.
- Maintenance & community: [Supermemory API + MCP](https://www.anchorterminal.com/tools/supermemory.md), 85/100 on maintenance & community, against 83 for the overall leader.
- Paying per call with no account (x402): [Honcho](https://www.anchorterminal.com/tools/honcho.md), accepts x402, $0.001 a call.
- A hosted MCP endpoint: [Zep](https://www.anchorterminal.com/tools/zep.md), remote MCP server, nothing to install.
- Self-hosting under an open licence: [Honcho](https://www.anchorterminal.com/tools/honcho.md), self-hosted, AGPL-3 licence. Also [Supermemory API + MCP](https://www.anchorterminal.com/tools/supermemory.md), self-hosted, MIT licence.

## How to choose

- Recall of early facts: Check whether facts from early sessions come back after many later ones, since a memory layer that forgets the first sessions can hand the agent stale context.
- Updates to changed preferences: Check that a changed preference replaces the old value rather than sitting beside it, because an agent that recalls both may act on the outdated one.
- Deletion that sticks: Check that a deleted fact stops appearing in recall and whether copies remain in logs or derived indexes, since a fact that returns after deletion is a privacy failure.
- Where memory is stored and encrypted: Check where the memory store is hosted, how it is encrypted and whether the vendor trains on what you send, because a user's history is personal data.

- How the benchmark tests 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.

## Each one in detail

### 1. Amazon Bedrock AgentCore Memory, A 79.4/100

Managed memory service for AI agents on AWS. It stores conversation events as short-term memory and extracts facts, preferences, summaries and episodes into searchable long-term records, through the AWS API, SDKs and an MCP server.

- Verdict: Access is IAM-controlled down to one namespace, with published per-second quotas for every operation and a `clientToken` on event writes. Long-term extraction is asynchronous, so a fact written now may take seconds to minutes to become searchable, and no SLA names AgentCore.
- Choose it for: Teams already on AWS that want per-user memory under IAM, KMS and Regional controls, with extraction run for them.
- Strength: IAM permissions per operation, resource-based policies and namespace condition keys, with OAuth sign-in and deny-by-default Cedar policies available through AgentCore Gateway
- Strength: Per-second quotas published for every Memory operation, such as 200 `CreateEvent` and 30 `RetrieveMemoryRecords` requests a second per account and Region
- Strength: `CreateEvent` takes a `clientToken`, so a retried write is ignored instead of stored twice
- Weakness: Long-term extraction is asynchronous. The docs say records appear within seconds to minutes after a write
- Weakness: No SLA names AgentCore. The Bedrock SLA of 4 October 2023 covers the Bedrock APIs for models
- Weakness: Built-in strategies use cross-Region inference, so event text can be processed in another Region of the same geography
- Price: $1 / GB · Auth: OAuth or key · x402: no · Where: hosted and local
- Full assessment: https://www.anchorterminal.com/tools/agentcore-memory.md

### 2. Zep, B 69.3/100

Hosted context and memory service built on a temporal graph of facts, relationships and source episodes.

- Verdict: Facts that new data invalidates keep the time they stopped being true. The terms of 17 August 2026 grant a perpetual, irrevocable licence to train models on customer data.
- Choose it for: Agents that must track how facts about a person or account change over time, inside a team that needs access policies and audit logs.
- Strength: Facts that new data invalidates keep the time they stopped being true
- Strength: API keys can carry ABAC policies, and audit and API logs record what each key did
- Strength: Published rate limits (600 a minute on Flex, 1,000 on Flex Plus) with 429, Retry-After and X-RateLimit headers
- Weakness: The terms of 17 August 2026 grant a perpetual, irrevocable licence to train models on customer data
- Weakness: First paid plan is $125 a month, and credits grow with episode size
- Weakness: The MCP server needs OAuth through a company identity provider, so a headless agent can't use it
- Price: $125 / mo · Auth: OAuth or key · x402: no · Where: hosted
- Full assessment: https://www.anchorterminal.com/tools/zep.md
- Against #1: https://www.anchorterminal.com/compare/agentcore-memory-vs-zep.md

### 3. Honcho, B 64.1/100

Memory API that models each participant (a peer) in a conversation.

- Verdict: Pay per call over x402 or MPP at agentcash.honcho.dev, with free read endpoints. No published rate limits, 429 guidance or SLA.
- Choose it for: Products that model the people in a conversation and want to ask questions about them, and for agents that must pay their own way.
- Strength: Pay per call over x402 or MPP at agentcash.honcho.dev, with free read endpoints
- Strength: Keys can be minted per workspace, peer or session, with an expiry, through the API
- Strength: Context retrieval listed as unlimited, with reasoning billed only when you call chat
- Weakness: No published rate limits, 429 guidance or SLA
- Weakness: Changelog entries carry no dates and the repo has no GitHub releases
- Weakness: The MCP tool list is sent on connect, not documented
- Price: $0.001 / call · Auth: OAuth or key · x402: yes · Where: hosted
- Full assessment: https://www.anchorterminal.com/tools/honcho.md
- Against #1: https://www.anchorterminal.com/compare/agentcore-memory-vs-honcho.md

### 4. Supermemory API + MCP, B 63.6/100

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.

- Verdict: 8-tool hosted MCP with OAuth and per-space read or write permission. No legal entity named in the terms or privacy policy.
- Choose it for: Apps that need user memory and document search over the same data, and for multi-tenant products that want a key per user.
- Strength: 8-tool hosted MCP with OAuth and per-space read or write permission
- Strength: Scoped keys limited to container tags, with expiry from 1 to 365 days
- Strength: Status page shows 100 per cent for the API and Console from July to October 2026
- Weakness: No legal entity named in the terms or privacy policy
- Weakness: No published rate limits or 429 guidance
- Weakness: Ingests web pages and PDFs with no prompt-injection guidance found
- Price: $19 / mo · Auth: OAuth or key · x402: no · Where: hosted
- Full assessment: https://www.anchorterminal.com/tools/supermemory.md
- Against #1: https://www.anchorterminal.com/compare/agentcore-memory-vs-supermemory.md

### 5. Mem0 Platform + MCP, C 56.1/100

Hosted memory layer that extracts facts from conversations and returns the relevant ones for a user, agent or run on later turns.

- Verdict: An agent can create its own Free account with `mem0 init --agent`, no email and no card. Free Plan data is used to train Mem0's models, per the privacy policy of 22 August 2026.
- Choose it for: Chat products that want per-user facts back with one search call and little setup.
- Strength: An agent can create its own Free account with `mem0 init --agent`, no email and no card
- Strength: Hosted MCP server with 11 tools, listed in the official MCP registry as io.github.mem0ai/mem0
- Strength: Public OpenAPI spec, llms.txt and Python and TypeScript SDKs, both released on 2026-09-25
- Weakness: Free Plan data is used to train Mem0's models, per the privacy policy of 22 August 2026
- Weakness: No published rate limits or 429 handling, and only 400 and 404 documented as errors
- Weakness: Retrieval caps are low below Pro, 1,000 a month free and 5,000 on Starter
- Price: $19 / mo · Auth: API key · x402: no · Where: hosted
- Full assessment: https://www.anchorterminal.com/tools/mem0.md
- Against #1: https://www.anchorterminal.com/compare/agentcore-memory-vs-mem0.md

### 6. Cognee, C 54.6/100

Open-source memory engine that turns documents, conversations and synced sources into a knowledge graph plus a vector index and answers queries over both.

- Verdict: Apache-2.0 library, REST server and MCP server, all self-hostable, with local models and no LLM key since 1.6.0. No status page, rate limits, SLA or SOC 2 for Cloud, and Cloud runs in AWS us-east-1 with no EU region.
- Choose it for: Agents whose memory has to include documents, wikis and chat tools as well as conversation, and for teams happy to self-host.
- Strength: Apache-2.0 library, REST server and MCP server, all self-hostable, with local models and no LLM key since 1.6.0
- Strength: 7-tool MCP server with search_tools and call_tool for reaching the rest on demand
- Strength: Public OpenAPI 3.1 file with 46 paths and error models
- Weakness: No status page, rate limits, SLA or SOC 2 for Cloud, and Cloud runs in AWS us-east-1 with no EU region
- Weakness: Cloud calls hung rather than failing when a tenant ran out of credit (August 2026), and the issue is still open
- Weakness: Billing docs and pricing page disagree on plans and the free allowance
- Price: $5 / mo · Auth: OAuth or key · x402: no · Where: hosted and local
- Full assessment: https://www.anchorterminal.com/tools/cognee.md
- Against #1: https://www.anchorterminal.com/compare/agentcore-memory-vs-cognee.md

### 7. Memory (MCP reference server), C 54.2/100

Knowledge-graph persistent memory reference server (entities, relations, observations) stored as JSONL at MEMORY_FILE_PATH.

- Verdict: Nine tools with typed schemas and accurate read, destructive and idempotent annotations. No pagination or limits. `read_graph` returns everything and `search_nodes` every match.
- Choose it for: A single local agent that wants a small, inspectable store of facts about people and projects.
- Strength: Nine tools with typed schemas and accurate read, destructive and idempotent annotations
- Strength: Plain JSONL storage at a path you choose, easy to back up, diff and edit by hand
- Strength: Writes are atomic since 2026.8.31, so an interrupted save can't truncate the file
- Weakness: No pagination or limits. `read_graph` returns everything and `search_nodes` every match
- Weakness: The published version can lose one of two writes made in the same turn. The fix is merged but unreleased
- Weakness: Search is case-insensitive substring matching, with no ranking or semantics
- Price: Free · OSS · Auth: None · x402: no · Where: local
- Full assessment: https://www.anchorterminal.com/tools/memory-reference-server.md
- Against #1: https://www.anchorterminal.com/compare/agentcore-memory-vs-memory-reference-server.md
- Disclosure: MCP started at Anthropic, which makes the Claude models our research agents and review panel run on (Anthropic donated it to the Agentic AI Foundation, a directed fund under the Linux Foundation, in December 2025), and this server is graded by the same checklist as every other listing.

### 8. Graphiti, D 53.3/100

Open-source Python framework from Zep that builds a temporal knowledge graph from chat messages, text and JSON.

- Verdict: Apache-2.0, with FalkorDB, Neo4j or Amazon Neptune as the store. Every add runs LLM extraction, so ingestion costs tokens and time.
- Choose it for: Teams that want Zep's temporal graph model on their own infrastructure and can run a graph database.
- Strength: Apache-2.0, with FalkorDB, Neo4j or Amazon Neptune as the store
- Strength: 13-tool MCP server over streamable HTTP or stdio, with a Docker Compose file
- Strength: Works with OpenAI, Anthropic, Gemini, Groq or a local OpenAI-compatible model
- Weakness: Every add runs LLM extraction, so ingestion costs tokens and time
- Weakness: The MCP HTTP endpoint has no authentication, and no tool carries readOnlyHint or destructiveHint
- Weakness: Still 0.x, and the last three PyPI releases have no GitHub release notes
- Price: Free · OSS · Auth: None · x402: no · Where: library
- Full assessment: https://www.anchorterminal.com/tools/graphiti.md
- Against #1: https://www.anchorterminal.com/compare/agentcore-memory-vs-graphiti.md

### 9. Hindsight, D 50.2/100

Memory engine for storing and retrieving information used by agents.

- Verdict: API keys support bank-level restrictions, expiry and child-key revocation. Retain ingestion costs $10 per million tokens.
- Choose it for: Agents that should form opinions and summaries from what they stored, such as long-running assistants or copilots that reflect on past sessions.
- Strength: Keys restricted to named banks, with expiry from an hour to a year and child-key revocation
- Strength: `Memory Defense` screens every retain, with regex redaction even in the open-source server
- Strength: Published per-token and per-call Cloud prices, with no monthly fee, and a published 99.9 per cent SLA
- Weakness: Retain at $10 per million tokens is the priciest ingestion in this category
- Weakness: 27 MCP tools per bank, with no subset or read-only option
- Weakness: No status page, no published rate limits and no llms.txt
- Price: $0.05 / call · Auth: OAuth or key · x402: no · Where: hosted
- Full assessment: https://www.anchorterminal.com/tools/hindsight.md
- Against #1: https://www.anchorterminal.com/compare/agentcore-memory-vs-hindsight.md

### 10. LangMem, D 47.9/100

LangMem is LangChain's open-source Python library for long-term agent memory. It extracts facts from conversations with an LLM, stores and searches them in a LangGraph store the owner runs, and includes two memory tools, message summarisation and prompt optimisation.

- Verdict: Two compact, typed memory tools and a background extraction manager work with any LangGraph store, under the MIT licence with nothing to buy. The newest PyPI release, 0.0.30, dates from 27 October 2025. The repository has no changelog, tags or test workflow, and 23 of 54 open issues have no reply.
- Choose it for: Teams already on LangGraph that want memory tools and background extraction over a store they run.
- Strength: MIT licence, installed with `pip install -U langmem`, with no account, key or fee of its own
- Strength: Two agent tools, `manage_memory` and `search_memory`, with typed inputs, an action enum and `limit`, `offset` and `filter` on search
- Strength: `actions_permitted` limits the manage tool to any subset of create, update and delete, and `create_memory_store_manager` leaves deletes off by default
- Weakness: The newest PyPI release, 0.0.30, is from 27 October 2025, and every commit to `src` on main is older than that
- Weakness: No changelog, GitHub releases or tags, and versions are still 0.0.x
- Weakness: The repository's two workflows deploy docs and publish to PyPI. Neither runs the tests
- Price: Free · OSS · Auth: None · x402: no · Where: local
- Full assessment: https://www.anchorterminal.com/tools/langmem.md
- Against #1: https://www.anchorterminal.com/compare/agentcore-memory-vs-langmem.md

## Head to head

- [Amazon Bedrock AgentCore Memory vs Zep](https://www.anchorterminal.com/compare/agentcore-memory-vs-zep.md)
- [Amazon Bedrock AgentCore Memory vs Honcho](https://www.anchorterminal.com/compare/agentcore-memory-vs-honcho.md)
- [Amazon Bedrock AgentCore Memory vs Supermemory API + MCP](https://www.anchorterminal.com/compare/agentcore-memory-vs-supermemory.md)
- [Amazon Bedrock AgentCore Memory vs Mem0 Platform + MCP](https://www.anchorterminal.com/compare/agentcore-memory-vs-mem0.md)
- [Honcho vs Zep](https://www.anchorterminal.com/compare/honcho-vs-zep.md)
- [Supermemory API + MCP vs Zep](https://www.anchorterminal.com/compare/supermemory-vs-zep.md)
- [Mem0 Platform + MCP vs Zep](https://www.anchorterminal.com/compare/mem0-vs-zep.md)
- [Honcho vs Supermemory API + MCP](https://www.anchorterminal.com/compare/honcho-vs-supermemory.md)
- [Honcho vs Mem0 Platform + MCP](https://www.anchorterminal.com/compare/honcho-vs-mem0.md)
- [Mem0 Platform + MCP vs Supermemory API + MCP](https://www.anchorterminal.com/compare/mem0-vs-supermemory.md)

## Questions

### What are the highest-rated memory layers for AI agents?

Amazon Bedrock AgentCore Memory has the highest benchmark score of the 10 ranked memory layers, 79.4 (A). Zep is second with 69.3 (B).

### How many memory layers are agent-ready?

1 of the 10 ranked here grade BB or better, the bar for agent-ready on the Anchor benchmark.

### Which memory layers accept x402 payments?

Honcho. An agent can pay these per call in USDC with no account.

### How is this list ranked?

By the Anchor benchmark score out of 100, a weighted mean of the scored categories minus deductions for negative events, from public evidence re-checked as vendors change. Listings cannot pay for a place. The latest assessment behind this page is from 8 October 2026.

## How this list is made

The order is the Anchor benchmark score, the same number as on each listing. Each listing is graded from public evidence against the benchmark checklist, and the picks are worked out from those grades, prices and facts. No listing pays for its place, and paid audits or listing help never change a score.