my-project-public

repository

loading code, commits, and activity

repositories

loading repo index

#1	---
#2	title: "Mem0 MCP"
#3	description: "Connect any AI client to Mem0 using Model Context Protocol in minutes"
#4	icon: "puzzle-piece"
#5	estimatedTime: "~5 minutes"
#6	---
#7
#8	<Info>
#9	Prerequisites
#10	- Mem0 Platform account ([Sign up here](https://app.mem0.ai))
#11	- API key ([Get one from dashboard](https://app.mem0.ai/settings/api-keys))
#12	- Python 3.10+, Docker, or Node.js 14+
#13	- An MCP-compatible client (Claude Desktop, Cursor, or custom agent)
#14	</Info>
#15
#16	## What is Mem0 MCP?
#17
#18	Mem0 MCP Server exposes Mem0's memory capabilities as MCP tools, letting AI agents decide when to save, search, or update information.
#19
#20	## Deployment Options
#21
#22	Choose from three deployment methods:
#23
#24	1. Python Package (Recommended) - Install locally with `uvx` for instant setup
#25	2. Docker Container - Isolated deployment with HTTP endpoint
#26	3. Smithery - Remote hosted service for managed deployments
#27
#28	## Available Tools
#29
#30	The MCP server exposes these memory tools to your AI client:
#31
#32	\| Tool \| Description \|
#33	\|------\|-------------\|
#34	\| `add_memory` \| Save text or conversation history for a user/agent \|
#35	\| `search_memories` \| Semantic search across existing memories with filters \|
#36	\| `get_memories` \| List memories with structured filters and pagination \|
#37	\| `get_memory` \| Retrieve one memory by its `memory_id` \|
#38	\| `update_memory` \| Overwrite a memory's text after confirming the ID \|
#39	\| `delete_memory` \| Delete a single memory by `memory_id` \|
#40	\| `delete_all_memories` \| Bulk delete all memories in scope \|
#41	\| `delete_entities` \| Delete a user/agent/app/run entity and its memories \|
#42	\| `list_entities` \| Enumerate users/agents/apps/runs stored in Mem0 \|
#43
#44	---
#45
#46	## Quickstart with Python (UVX)
#47
#48	<Steps>
#49	<Step title="Install the MCP Server">
#50	```bash
#51	uv pip install mem0-mcp-server
#52	```
#53	</Step>
#54
#55	<Step title="Configure your MCP client">
#56	Add this to your MCP client (e.g., Claude Desktop):
#57
#58	```json
#59	{
#60	"mcpServers": {
#61	"mem0": {
#62	"command": "uvx",
#63	"args": ["mem0-mcp-server"],
#64	"env": {
#65	"MEM0_API_KEY": "m0-...",
#66	"MEM0_DEFAULT_USER_ID": "your-handle"
#67	}
#68	}
#69	}
#70	}
#71	```
#72
#73	Set your environment variables:
#74
#75	```bash
#76	export MEM0_API_KEY="m0-..."
#77	export MEM0_DEFAULT_USER_ID="your-handle"
#78	```
#79	</Step>
#80
#81	<Step title="Test with the Python agent">
#82	```bash
#83	# Clone the mem0-mcp repository
#84	git clone https://github.com/mem0ai/mem0-mcp.git
#85	cd mem0-mcp
#86
#87	# Set your API keys
#88	export MEM0_API_KEY="m0-..."
#89	export OPENAI_API_KEY="sk-openai-..."
#90
#91	# Run the interactive agent
#92	python example/pydantic_ai_repl.py
#93	```
#94
#95	Sample Interactions:
#96
#97	```
#98	User: Remember that I love tiramisu
#99	Agent: Got it! I've saved that you love tiramisu.
#100
#101	User: What do you know about my food preferences?
#102	Agent: Based on your memories, you love tiramisu.
#103
#104	User: Update my project: the mobile app is now 80% complete
#105	Agent: Updated your project status successfully.
#106	```
#107	</Step>
#108
#109	<Step title="Verify the setup">
#110	Your AI client can now:
#111	- Automatically save information with `add_memory`
#112	- Search memories with `search_memories`
#113	- Update memories with `update_memory`
#114	- Delete memories with `delete_memory`
#115
#116	<Info icon="check">
#117	If you get "Connection failed", ensure your API key is valid and the server is running.
#118	</Info>
#119	</Step>
#120	</Steps>
#121
#122	---
#123
#124	## Quickstart with Docker
#125
#126	<Steps>
#127	<Step title="Build the Docker image">
#128	```bash
#129	docker build -t mem0-mcp-server https://github.com/mem0ai/mem0-mcp.git
#130	```
#131	</Step>
#132
#133	<Step title="Run the container">
#134	```bash
#135	docker run --rm -d \
#136	--name mem0-mcp \
#137	-e MEM0_API_KEY="m0-..." \
#138	-p 8080:8081 \
#139	mem0-mcp-server
#140	```
#141	</Step>
#142
#143	<Step title="Configure your client for HTTP">
#144	For clients that connect via HTTP (instead of stdio):
#145
#146	```json
#147	{
#148	"mcpServers": {
#149	"mem0-docker": {
#150	"command": "curl",
#151	"args": ["-X", "POST", "http://localhost:8080/mcp", "--data-binary", "@-"],
#152	"env": {
#153	"MEM0_API_KEY": "m0-..."
#154	}
#155	}
#156	}
#157	}
#158	```
#159	</Step>
#160
#161	<Step title="Verify the setup">
#162	```bash
#163	# Check container logs
#164	docker logs mem0-mcp
#165
#166	# Test HTTP endpoint
#167	curl http://localhost:8080/health
#168	```
#169
#170	<Info icon="check">
#171	The container should start successfully and respond to HTTP requests. If port 8080 is occupied, change it with `-p 8081:8081`.
#172	</Info>
#173	</Step>
#174	</Steps>
#175
#176	---
#177
#178	## Quickstart with Smithery (Hosted)
#179
#180	For the simplest integration, use Smithery's hosted Mem0 MCP server - no installation required.
#181
#182	Example: One-click setup in Cursor
#183
#184	1. Visit [smithery.ai/server/@mem0ai/mem0-memory-mcp](https://smithery.ai/server/@mem0ai/mem0-memory-mcp) and select Cursor as your client
#185
#186	![Smithery Mem0 MCP Configuration](/images/smithery-mem0-mcp.png)
#187
#188	2. Open Cursor → Settings → MCP
#189	3. Click `mem0-mcp` → Initiate authorization
#190	4. Configure Smithery with your environment:
#191	- `MEM0_API_KEY`: Your Mem0 API key
#192	- `MEM0_DEFAULT_USER_ID`: Your user ID
#193	- `MEM0_ENABLE_GRAPH_DEFAULT`: Optional, set to `true` for graph memories
#194	5. Return to Cursor settings and wait for tools to load
#195	6. Start chatting with Cursor and begin storing preferences
#196
#197	For other clients:
#198	Visit [smithery.ai/server/@mem0ai/mem0-memory-mcp](https://smithery.ai/server/@mem0ai/mem0-memory-mcp) to connect any MCP-compatible client with your Mem0 credentials.
#199
#200	---
#201
#202	## Quick Recovery
#203
#204	- "uvx command not found" → Install with `pip install uv` or use `pip install mem0-mcp-server` instead. Make sure your Python environment has `uv` installed (or system-wide).
#205	- "Connection refused" → Check that the server is running and the correct port is configured
#206	- "Invalid API key" → Get a new key from [Mem0 Dashboard](https://app.mem0.ai/settings/api-keys)
#207	- "Permission denied" → Ensure Docker has access to bind ports (try with `sudo` on Linux)
#208
#209	---
#210
#211	## Next Steps
#212
#213	<CardGroup cols={2}>
#214	<Card
#215	title="MCP Integration Feature"
#216	description="Learn about MCP configuration options and advanced patterns"
#217	icon="plug"
#218	href="/platform/features/mcp-integration"
#219	/>
#220	<Card
#221	title="Gemini 3 with Mem0 MCP"
#222	description="See how to integrate Gemini 3 with Mem0 MCP server"
#223	icon="book-open"
#224	href="/cookbooks/frameworks/gemini-3-with-mem0-mcp"
#225	/>
#226	</CardGroup>
#227
#228	## Additional Resources
#229
#230	- [Mem0 MCP Repository](https://github.com/mem0ai/mem0-mcp) - Source code and examples
#231	- [Platform Quickstart](/platform/quickstart) - Direct API integration guide
#232	- [MCP Specification](https://modelcontextprotocol.io) - Learn about MCP protocol

z6Mkq5mY3JWtxoxUobWcfNHm7AkRubgSWEZTkBVqZXJviFZ5/my-project-public