openmemory

Name: openmemory
Author: caviraoss

bycaviraoss· 1 MCP server

Stars

4.4k

Forks

494

How to install openmemory?

caviraoss/openmemory

$git clone https://github.com/caviraoss/openmemory

Installs into the current project.

›Prefer a prompt? Paste this to your agent

Install & use

Install openmemory by running `git clone https://github.com/caviraoss/openmemory`, then use it for the current task and follow its documentation at https://github.com/caviraoss/openmemory.

Files · 1

View on GitHub

README.md

1	# 🚧 This project is currently being rewritten.
2
3	Expect breaking changes and potential bugs.
4	### Contributers needed
5	To contribute, visit https://github.com/CaviraOSS/OpenMemory/tree/rewrite branch.
6	If you find an issue, please open a GitHub issue with details so it can be tracked and resolved.
7	## OpenMemory
8
9	> Real long-term memory for AI agents. Not RAG. Not a vector DB. Self-hosted, Python + Node.
10
11	[![VS Code Extension](https://img.shields.io/badge/VS%20Code-Extension-007ACC?logo=visualstudiocode)](https://marketplace.visualstudio.com/items?itemName=Nullure.openmemory-vscode)
12	[![Discord](https://img.shields.io/discord/1300368230320697404?label=Discord)](https://discord.gg/93M9XSuEj6)
13	[![PyPI](https://img.shields.io/pypi/v/openmemory-py.svg)](https://pypi.org/project/openmemory-py/)
14	[![npm](https://img.shields.io/npm/v/openmemory-js.svg)](https://www.npmjs.com/package/openmemory-js)
15	[![License](https://img.shields.io/github/license/CaviraOSS/OpenMemory)](LICENSE)
16
17	![OpenMemory demo](.github/openmemory.gif)
18
19	OpenMemory is a cognitive memory engine for LLMs and agents.
20
21	- 🧠 Real long-term memory (not just embeddings in a table)
22	- 💾 Self-hosted, local-first (SQLite / Postgres)
23	- 🐍 Python + 🟦 Node SDKs
24	- 🧩 Integrations: LangChain, CrewAI, AutoGen, Streamlit, MCP, VS Code
25	- 📥 Sources: GitHub, Notion, Google Drive, OneDrive, Web Crawler
26	- 🔍 Explainable traces (see why something was recalled)
27
28	Your model stays stateless. Your app stops being amnesiac.
29
30	---
31
32	## 1. TL;DR – Use It in 10 Seconds
33
34	### 🐍 Python (local-first)
35
36	Install:
37
38	```bash
39	pip install openmemory-py
40	```
41
42	Use:
43
44	```python
45	from openmemory.client import Memory
46
47	mem = Memory()
48	mem.add("user prefers dark mode", user_id="u1")
49	results = mem.search("preferences", user_id="u1")
50	await mem.delete("memory_id")
51	```
52
53	> Note: `add`, `search`, `get`, `delete` are async. Use `await` in async contexts.
54
55	#### 🔗 OpenAI
56
57	```python
58	mem = Memory()
59	client = mem.openai.register(OpenAI(), user_id="u1")
60	resp = client.chat.completions.create(...)
61	```
62
63	#### 🧱 LangChain
64
65	```python
66	from openmemory.integrations.langchain import OpenMemoryChatMessageHistory
67
68	history = OpenMemoryChatMessageHistory(memory=mem, user_id="u1")
69	```
70
71	#### 🤝 CrewAI / AutoGen / Streamlit
72
73	OpenMemory is designed to sit behind agent frameworks and UIs:
74
75	- Crew-style agents: use `Memory` as a shared long-term store
76	- AutoGen-style orchestrations: store dialog + tool calls as episodic memory
77	- Streamlit apps: give each user a persistent memory by `user_id`
78
79	See the integrations section in the docs for concrete patterns.
80
81	---
82
83	### 🟦 Node / JavaScript (local-first)
84
85	Install:
86
87	```bash
88	npm install openmemory-js
89	```
90
91	Use:
92
93	```ts
94	import { Memory } from "openmemory-js"
95
96	const mem = new Memory()
97	await mem.add("user likes spicy food", { user_id: "u1" })
98	const results = await mem.search("food?", { user_id: "u1" })
99	await mem.delete("memory_id")
100	```
101
102	Drop this into:
103
104	- Node backends
105	- CLIs
106	- local tools
107	- anything that needs durable memory without running a separate service.
108
109	---
110
111	### 📥 Connectors
112
113	Ingest data from external sources directly into memory:
114
115	```python
116	# python
117	github = mem.source("github")
118	await github.connect(token="ghp_...")
119	await github.ingest_all(repo="owner/repo")
120	```
121
122	```ts
123	// javascript
124	const github = await mem.source("github")
125	await github.connect({ token: "ghp_..." })
126	await github.ingest_all({ repo: "owner/repo" })
127	```
128
129	Available connectors: `github`, `notion`, `google_drive`, `google_sheets`, `google_slides`, `onedrive`, `web_crawler`
130
131	---
132
133	## 2. Modes: SDKs, Server, MCP
134
135	OpenMemory can run inside your app or as a central service.
136
137	### 2.1 Python SDK
138
139	- ✅ Local SQLite by default
140	- ✅ Supports external DBs (via config)
141	- ✅ Great fit for LangChain / LangGraph / CrewAI / notebooks
142
143	Docs: https://openmemory.cavira.app/docs/sdks/python
144
145	---
146
147	### 2.2 Node SDK
148
149	- Same cognitive model as Python
150	- Ideal for JS/TS applications
151	- Can either run fully local or talk to a central backend
152
153	Docs: h

6 picks

Type

openmemory

bycaviraoss· 1 MCP server

Stars

4.4k

Forks

494

How to install openmemory?

caviraoss/openmemory

$git clone https://github.com/caviraoss/openmemory

Installs into the current project.

›Prefer a prompt? Paste this to your agent

Install & use

Install openmemory by running `git clone https://github.com/caviraoss/openmemory`, then use it for the current task and follow its documentation at https://github.com/caviraoss/openmemory.

Files · 1

View on GitHub

README.md

1	# 🚧 This project is currently being rewritten.
2
3	Expect breaking changes and potential bugs.
4	### Contributers needed
5	To contribute, visit https://github.com/CaviraOSS/OpenMemory/tree/rewrite branch.
6	If you find an issue, please open a GitHub issue with details so it can be tracked and resolved.
7	## OpenMemory
8
9	> Real long-term memory for AI agents. Not RAG. Not a vector DB. Self-hosted, Python + Node.
10
11	[![VS Code Extension](https://img.shields.io/badge/VS%20Code-Extension-007ACC?logo=visualstudiocode)](https://marketplace.visualstudio.com/items?itemName=Nullure.openmemory-vscode)
12	[![Discord](https://img.shields.io/discord/1300368230320697404?label=Discord)](https://discord.gg/93M9XSuEj6)
13	[![PyPI](https://img.shields.io/pypi/v/openmemory-py.svg)](https://pypi.org/project/openmemory-py/)
14	[![npm](https://img.shields.io/npm/v/openmemory-js.svg)](https://www.npmjs.com/package/openmemory-js)
15	[![License](https://img.shields.io/github/license/CaviraOSS/OpenMemory)](LICENSE)
16
17	![OpenMemory demo](.github/openmemory.gif)
18
19	OpenMemory is a cognitive memory engine for LLMs and agents.
20
21	- 🧠 Real long-term memory (not just embeddings in a table)
22	- 💾 Self-hosted, local-first (SQLite / Postgres)
23	- 🐍 Python + 🟦 Node SDKs
24	- 🧩 Integrations: LangChain, CrewAI, AutoGen, Streamlit, MCP, VS Code
25	- 📥 Sources: GitHub, Notion, Google Drive, OneDrive, Web Crawler
26	- 🔍 Explainable traces (see why something was recalled)
27
28	Your model stays stateless. Your app stops being amnesiac.
29
30	---
31
32	## 1. TL;DR – Use It in 10 Seconds
33
34	### 🐍 Python (local-first)
35
36	Install:
37
38	```bash
39	pip install openmemory-py
40	```
41
42	Use:
43
44	```python
45	from openmemory.client import Memory
46
47	mem = Memory()
48	mem.add("user prefers dark mode", user_id="u1")
49	results = mem.search("preferences", user_id="u1")
50	await mem.delete("memory_id")
51	```
52
53	> Note: `add`, `search`, `get`, `delete` are async. Use `await` in async contexts.
54
55	#### 🔗 OpenAI
56
57	```python
58	mem = Memory()
59	client = mem.openai.register(OpenAI(), user_id="u1")
60	resp = client.chat.completions.create(...)
61	```
62
63	#### 🧱 LangChain
64
65	```python
66	from openmemory.integrations.langchain import OpenMemoryChatMessageHistory
67
68	history = OpenMemoryChatMessageHistory(memory=mem, user_id="u1")
69	```
70
71	#### 🤝 CrewAI / AutoGen / Streamlit
72
73	OpenMemory is designed to sit behind agent frameworks and UIs:
74
75	- Crew-style agents: use `Memory` as a shared long-term store
76	- AutoGen-style orchestrations: store dialog + tool calls as episodic memory
77	- Streamlit apps: give each user a persistent memory by `user_id`
78
79	See the integrations section in the docs for concrete patterns.
80
81	---
82
83	### 🟦 Node / JavaScript (local-first)
84
85	Install:
86
87	```bash
88	npm install openmemory-js
89	```
90
91	Use:
92
93	```ts
94	import { Memory } from "openmemory-js"
95
96	const mem = new Memory()
97	await mem.add("user likes spicy food", { user_id: "u1" })
98	const results = await mem.search("food?", { user_id: "u1" })
99	await mem.delete("memory_id")
100	```
101
102	Drop this into:
103
104	- Node backends
105	- CLIs
106	- local tools
107	- anything that needs durable memory without running a separate service.
108
109	---
110
111	### 📥 Connectors
112
113	Ingest data from external sources directly into memory:
114
115	```python
116	# python
117	github = mem.source("github")
118	await github.connect(token="ghp_...")
119	await github.ingest_all(repo="owner/repo")
120	```
121
122	```ts
123	// javascript
124	const github = await mem.source("github")
125	await github.connect({ token: "ghp_..." })
126	await github.ingest_all({ repo: "owner/repo" })
127	```
128
129	Available connectors: `github`, `notion`, `google_drive`, `google_sheets`, `google_slides`, `onedrive`, `web_crawler`
130
131	---
132
133	## 2. Modes: SDKs, Server, MCP
134
135	OpenMemory can run inside your app or as a central service.
136
137	### 2.1 Python SDK
138
139	- ✅ Local SQLite by default
140	- ✅ Supports external DBs (via config)
141	- ✅ Great fit for LangChain / LangGraph / CrewAI / notebooks
142
143	Docs: https://openmemory.cavira.app/docs/sdks/python
144
145	---
146
147	### 2.2 Node SDK
148
149	- Same cognitive model as Python
150	- Ideal for JS/TS applications
151	- Can either run fully local or talk to a central backend
152
153	Docs: h

6 picks

Type

openmemory

How to install openmemory?

Related

openmemory

How to install openmemory?

Related