Back to vendors
M

Mem0

Also known as: mem0, mem-zero, mem0ai

Visit site
Entry priceFrom $19/mo. A free tier covers 10k add and 1k retrieval requests a month, and the open source build is free.Full pricing detail

Open source memory layer for AI agents that extracts facts from interactions and returns relevant memories on later calls, scoped per user, agent, app and run, via API, SDKs, CLI and MCP.

Mem0 is a memory layer for AI agents and assistants. Language models are stateless, so every call starts from a blank slate. After each interaction Mem0 extracts the durable facts, such as a user's preferences, decisions and prior context, and stores them. On the next call it returns only the relevant ones to add to the prompt, so agents remember across sessions without replaying whole transcripts. It is open source under Apache 2.0 and also sold as the managed Mem0 Platform.

Memories are scoped by user, agent, application and run, so one agent can keep separate context per customer and short lived sessions can expire on their own. Memories can be updated, given an expiry date or deleted per user. Retrieval combines semantic search, keyword matching and entity linking, with optional decay of stale memories and a consolidation feature called Dream on higher plans. Mem0 works with LangGraph, CrewAI, the OpenAI Agents SDK, Google ADK, Mastra, Claude Code, Cursor and many other frameworks and tools. Developers reach it through a REST API, Python and TypeScript SDKs, a CLI and an MCP server, and an agent can mint its own API key from the command line.

The open source edition runs on the customer's own infrastructure with their choice of LLM, embedder and vector store, and Enterprise adds on premises deployment, audit logs and SSO. Mem0 holds SOC 2 Type I and is HIPAA and GDPR ready. The Platform has a free Hobby tier with 10,000 add and 1,000 retrieval requests a month, Starter at $19 a month, Pro at $249 a month with graph memory and Dream, and a custom Enterprise tier, with usage-based pricing available.

Vendor details

Canonical URL

https://mem0.ai

Category

Agent infrastructure

Subcategory

Agent memory

Funding status

Founded by Taranjeet Singh (CEO) and Deshraj Yadav (CTO), the team behind Embedchain, with around $24M raised. The open source project has roughly 48,000 GitHub stars. Independent.

Company status

independent

Use cases & customers

Primary use cases

agent memorypersonalizationcross session contexttoken cost reductionconversational recall

Target customers

developersenterprise

Deployment options

SaaSself-hostedon-prem

Integrations

Mem0 is framework agnostic, with an add and search API and SDKs. It integrates with LangChain, LangGraph, CrewAI and AWS Strands Agents, and it can run as an MCP server. Self hosting pairs a vector store such as Qdrant with Postgres, and managed backends include Amazon ElastiCache for Valkey and Neptune Analytics for graph memory.

In practice

Your support bot asks returning users the same questions every session. You add Mem0 keyed by user, so it stores each user's preferences and recalls them next time without replaying the whole transcript.

Your token bill is high because you stuff full chat history into every prompt. Mem0 extracts and retrieves only the relevant memories, cutting context size and cost while keeping the agent's recall.

You need agent memory that stays in your environment for compliance. You self host Mem0 alongside your own Qdrant and Postgres, or run it on premises on the Enterprise plan, keeping every stored memory inside your boundary.

Agentic Index coverage score

6.0 / 14 capabilities · 43%

Integrations & Tool Calling Partial

Mem0 ships native integrations that wire memory into more than twenty agent frameworks and coding tools, among them LangGraph, CrewAI, the OpenAI Agents SDK, Google ADK, Mastra, Claude Code and Cursor, plus n8n and Zapier, and webhooks notify other systems when memories change. These connect memory to agents, and Mem0 itself does not read from or write to other business systems on an agent's behalf.

SourceMem0, docs.mem0.ai integrations and platform/features/webhooks (indexed in llms.txt)read 2026-09-21

Workflow Orchestration Not documented

Mem0 provides a memory layer that agent frameworks call, and it does not orchestrate agent workflows, sequencing or branching itself.

Sourcemem0.airead 2026-09-21

Knowledge Grounding & RAG Not documented

Retrieval runs over the memory store Mem0 extracts from interactions, and no separate document index or knowledge base over the customer's own content is offered alongside it.

SourceMem0, docs.mem0.ai llms.txt and platform/features/entity-scoped-memoryread 2026-09-21

Human Oversight & Guardrails Not documented

The product stores and retrieves memories and offers no human review, approval workflow or content guardrails.

Sourcemem0.airead 2026-09-21

Security, Identity & Governance Full

Enterprise adds SSO and audit logs, memories are partitioned per user, agent and app to keep tenants separate, and the self hosted server runs with its own authentication and API keys. A SOC 2 Type I badge is displayed on the site, which links to a Trust Center.

SourceMem0, mem0.ai/pricing and docs.mem0.ai platform/features/entity-scoped-memoryread 2026-09-21

Observability & Auditability Partial

Pro adds advanced analytics and Enterprise adds audit logs of memory reads and writes, and memories can be inspected by user, agent, app and run in the dashboard, giving a record of what the agent stored and recalled. Mem0 sees only the memory calls made to it, not the steps and tool calls in the rest of the agent's run.

SourceMem0, mem0.ai/pricing and docs.mem0.ai platform/features/entity-scoped-memoryread 2026-09-21

Memory & State Persistence Full

Mem0 extracts facts from each interaction into a memory store the agent searches as context on later calls, scoped by user_id, agent_id, app_id and run_id so each customer, agent or session keeps its own memory; run-scoped memories expire independently, memories can be updated, given an expiry date or deleted for one user without touching others, and decay and consolidation keep stale facts from surfacing.

SourceMem0, docs.mem0.ai platform/features/entity-scoped-memory and llms.txtread 2026-09-21

Deployment & Data Residency Full

Customers can run Mem0 as the managed Platform, as open-source software self-hosted with the bundled REST server and dashboard on the customer's own vector store and Postgres, or as an on-premises deployment on the Enterprise plan.

SourceMem0, mem0.ai/pricing and docs.mem0.ai open-source/setup (indexed in llms.txt)read 2026-09-21

Prebuilt Agents, Templates & Packs Not documented

Example integrations are provided, but no prebuilt agents, templates or installable packs.

Sourcemem0.airead 2026-09-21

Triggers & Channel Coverage Not documented

The product is an API that agents call, and it has no triggers, scheduling or conversational channels of its own.

Sourcemem0.airead 2026-09-21

Model Flexibility & Routing Full

The open-source Memory class is configured with the customer's choice of LLM, embedder, vector store and reranker, and Mem0 works with AWS Bedrock and OpenAI-compatible endpoints, so the customer controls which model extracts and ranks memories. Mem0 does not route requests between models itself.

SourceMem0, docs.mem0.ai open-source/configuration (indexed in llms.txt) and github.com/mem0ai/mem0read 2026-09-21

APIs, SDKs & MCP Extensibility Full

The Platform exposes a REST API for adding, searching, updating, exporting and deleting memories, with Python and TypeScript SDK clients, a CLI and an MCP server, and the open-source edition ships a self-hosted REST API server.

SourceMem0, docs.mem0.ai platform/features/entity-scoped-memory and llms.txtread 2026-09-21

Testing, Debugging & Optimization Not documented

An evaluation framework and benchmark results are published for Mem0's own memory algorithm. Customers get no scored test cases, harness or quality gate for evaluating their own agents.

SourceMem0, docs.mem0.ai core-concepts/memory-evaluation (indexed in llms.txt) and github.com/mem0ai/mem0read 2026-09-21

Browser & Computer Use Not documented

Mem0 is a memory layer and does not automate browsers or computers, and its browser extension only captures memories.

Sourcemem0.airead 2026-09-21

The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded

Product updates

·Memory / statePartially Verified

Mem0 introduced a multi-signal retrieval redesign in its open-source memory algorithm. The update replaces external graph store support with built-in entity linking, running semantic similarity, BM25 keyword matching, and entity matching in parallel before fusing them into a single result score.

Bears on: Memory / state

View source
View all Mem0 updates →Tracked since Jul 2026 · Verified from public vendor sources

Pricing

From $19/mo. A free tier covers 10k add and 1k retrieval requests a month, and the open source build is free.

Subscription tiers metered by memory add and retrieval requests per month and by number of projects. Usage based pricing is also available.

Free tier

Included quota

Hobby free (10,000 add, 1,000 retrieval requests/mo, 1 project). Starter $19/mo (50,000 add, 5,000 retrieval, 1 project). Pro $249/mo (500,000 add, 50,000 retrieval, unlimited projects, private Slack, advanced analytics, graph memory entity linking, Dream consolidation). Enterprise custom (unlimited, SLA, on premises deployment, audit logs, custom integrations, SSO).

What is public

Self serve tier pricing is published (Hobby free, Starter $19, Pro $249), with add, retrieval and project limits. Enterprise is custom, and usage based pricing is available.

Billing mechanics

Subscription tiers metered by monthly add (write) and retrieval (read) requests and project count. Each add request triggers an LLM call to extract memories. Open source self hosting has no request metering. Enterprise adds on premises deployment and an SLA under custom pricing.

Cost watchouts

Each memory add triggers a paid LLM call, so chat workloads with many turns inflate both request counts and the underlying model cost. Self hosting moves spend to the customer's own vector store and Postgres infrastructure.

Variable cost rationale

Memory add and retrieval requests are metered against a monthly quota per tier, along with the number of projects. Request volume grows with conversation and agent traffic, and each add runs an LLM call to extract facts, so workloads heavy in writes use up quota fastest. Tier quotas cap the monthly volume, and going past one means moving up a tier or onto usage based pricing. Self hosting the Apache 2.0 open source build has no request metering and moves the cost to the customer's own infrastructure.

Additional watchouts

Add requests are metered and each involves an LLM call, so chat workloads heavy in writes consume quota and cost quickly. Graph memory and Dream are on Pro and above. On premises deployment, audit logs and SSO are Enterprise only.

Overage / add-ons

Plans cap monthly add and retrieval requests by tier. Exceeding a tier's quota means upgrading, and Mem0 also offers usage based pricing for needs that do not fit a tier. Self hosting the open source build carries no request metering.

Sales call required

No, self serve available

Free / trial

Free Hobby tier with 10,000 add and 1,000 retrieval requests a month, 1 project and community support. The Apache 2.0 open source build is free to self host.

Lowest paid plan

Starter at $19 per month, with 50,000 add and 5,000 retrieval requests a month

Commercial notes

Adoption is led by open source and aimed at developers first, with a free tier and a free Apache 2.0 self hosting path. It scales to managed tiers and an Enterprise plan for on premises deployment, SSO and audit logs.

Key ambiguities

Enterprise dollar pricing and usage based per request rates above plan quotas are not public.

Cancellation / refund

Hobby, Starter and Pro are self serve subscriptions, and Enterprise terms are set by contract. No cancellation or refund terms are published.

Support SLA / resale

Community support on Hobby and Starter, private Slack on Pro, and private Slack with an SLA on Enterprise.

Missing data

Enterprise dollar pricing and the exact usage based per request rates are not public.

Agentic Index verified 2026-09-21

Alternatives to Mem0

The closest documented capability profiles to Mem0 among agent infrastructure platforms tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.

  • Zep7.0 / 14Adds documented Knowledge Grounding & RAGMem0 vs Zep →
  • Cognee7.0 / 14Adds documented Workflow Orchestration and Knowledge Grounding & RAGMem0 vs Cognee →
  • Hyperspell7.0 / 14Adds documented Knowledge Grounding & RAG and Testing, Debugging & Optimization
  • Deepgram7.5 / 14Adds documented Workflow Orchestration and Human Oversight & Guardrails, among others
  • Rime3.5 / 14A lighter documented profile than Mem0
  • Speechmatics3.5 / 14A lighter documented profile than Mem0

Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded

Contact us

Found a vendor we missed? Have feedback on the index? We'd love to hear from you.