my-project-public

repository

loading code, commits, and activity

repositories

loading repo index

#1	# mem0-ts
#2
#3	A TypeScript implementation of the mem0 memory system, using OpenAI for embeddings and completions.
#4
#5	## Features
#6
#7	- Memory storage and retrieval using vector embeddings
#8	- Fact extraction from text using GPT-4
#9	- SQLite-based history tracking
#10	- Optional graph-based memory relationships
#11	- TypeScript type safety
#12	- Built-in OpenAI integration with default configuration
#13	- In-memory vector store implementation
#14	- Extensible architecture with interfaces for custom implementations
#15
#16	## Installation
#17
#18	1. Clone the repository:
#19
#20	```bash
#21	git clone <repository-url>
#22	cd mem0-ts
#23	```
#24
#25	2. Install dependencies:
#26
#27	```bash
#28	npm install
#29	```
#30
#31	3. Set up environment variables:
#32
#33	```bash
#34	cp .env.example .env
#35	# Edit .env with your OpenAI API key
#36	```
#37
#38	4. Build the project:
#39
#40	```bash
#41	npm run build
#42	```
#43
#44	## Usage
#45
#46	### Basic Example
#47
#48	```typescript
#49	import { Memory } from "mem0-ts";
#50
#51	// Create a memory instance with default OpenAI configuration
#52	const memory = new Memory();
#53
#54	// Or with minimal configuration (only API key)
#55	const memory = new Memory({
#56	embedder: {
#57	config: {
#58	apiKey: process.env.OPENAI_API_KEY,
#59	},
#60	},
#61	llm: {
#62	config: {
#63	apiKey: process.env.OPENAI_API_KEY,
#64	},
#65	},
#66	});
#67
#68	// Or with custom configuration
#69	const memory = new Memory({
#70	embedder: {
#71	provider: "openai",
#72	config: {
#73	apiKey: process.env.OPENAI_API_KEY,
#74	model: "text-embedding-3-small",
#75	},
#76	},
#77	vectorStore: {
#78	provider: "memory",
#79	config: {
#80	collectionName: "custom-memories",
#81	},
#82	},
#83	llm: {
#84	provider: "openai",
#85	config: {
#86	apiKey: process.env.OPENAI_API_KEY,
#87	model: "gpt-4-turbo-preview",
#88	},
#89	},
#90	});
#91
#92	// Add a memory
#93	await memory.add("The sky is blue", "user123");
#94
#95	// Search memories
#96	const results = await memory.search("What color is the sky?", "user123");
#97	```
#98
#99	### Default Configuration
#100
#101	The memory system comes with sensible defaults:
#102
#103	- OpenAI embeddings with `text-embedding-3-small` model
#104	- In-memory vector store
#105	- OpenAI GPT-4 Turbo for LLM operations
#106	- SQLite for history tracking
#107
#108	You only need to provide API keys - all other settings are optional.
#109
#110	### Methods
#111
#112	- `add(messages: string \| Message[], userId?: string, ...): Promise<SearchResult>`
#113	- `search(query: string, userId?: string, ...): Promise<SearchResult>`
#114	- `get(memoryId: string): Promise<MemoryItem \| null>`
#115	- `update(memoryId: string, data: string): Promise<{ message: string }>`
#116	- `delete(memoryId: string): Promise<{ message: string }>`
#117	- `deleteAll(userId?: string, ...): Promise<{ message: string }>`
#118	- `history(memoryId: string): Promise<any[]>`
#119	- `reset(): Promise<void>`
#120
#121	### Try the Example
#122
#123	We provide a comprehensive example in `examples/basic.ts` that demonstrates all the features including:
#124
#125	- Default configuration usage
#126	- In-memory vector store
#127	- PGVector store (with PostgreSQL)
#128	- Qdrant vector store
#129	- Redis vector store
#130	- Memory operations (add, search, update, delete)
#131
#132	To run the example:
#133
#134	```bash
#135	npm run example
#136	```
#137
#138	You can use this example as a template and modify it according to your needs. The example includes:
#139
#140	- Different vector store configurations
#141	- Various memory operations
#142	- Error handling
#143	- Environment variable usage
#144
#145	## Development
#146
#147	1. Build the project:
#148
#149	```bash
#150	npm run build
#151	```
#152
#153	2. Clean build files:
#154
#155	```bash
#156	npm run clean
#157	```
#158
#159	## Extending
#160
#161	The system is designed to be extensible. You can implement your own:
#162
#163	- Embedders by implementing the `Embedder` interface
#164	- Vector stores by implementing the `VectorStore` interface
#165	- Language models by implementing the `LLM` interface
#166
#167	## License
#168
#169	MIT
#170
#171	## Contributing
#172
#173	1. Fork the repository
#174	2. Create your feature branch
#175	3. Commit your changes
#176	4. Push to the branch
#177	5. Create a new Pull Request
#178

z6Mkq5mY3JWtxoxUobWcfNHm7AkRubgSWEZTkBVqZXJviFZ5/my-project-public