my-project-public

repository

loading code, commits, and activity

repositories

loading repo index

#1	---
#2	title: "Configure the OSS Stack"
#3	description: "Wire up Mem0 OSS with your preferred LLM, vector store, embedder, and reranker."
#4	icon: "sliders"
#5	---
#6
#7	# Configure Mem0 OSS Components
#8
#9	<Info>
#10	Prerequisites
#11	- Python 3.10+ with `pip` available
#12	- Running vector database (e.g., Qdrant, Postgres + pgvector) or access credentials for a managed store
#13	- API keys for your chosen LLM, embedder, and reranker providers
#14	</Info>
#15
#16	<Tip>
#17	Start from the <Link href="/open-source/python-quickstart">Python quickstart</Link> if you still need the base CLI and repository.
#18	</Tip>
#19
#20	## Install dependencies
#21
#22	<Tabs>
#23	<Tab title="Python">
#24	<Steps>
#25	<Step title="Install Mem0 OSS">
#26	```bash
#27	pip install mem0ai
#28	```
#29	</Step>
#30	<Step title="Add provider SDKs (example: Qdrant + OpenAI)">
#31	```bash
#32	pip install qdrant-client openai
#33	```
#34	</Step>
#35	</Steps>
#36	</Tab>
#37	<Tab title="Docker Compose">
#38	<Steps>
#39	<Step title="Clone the repo and copy the compose file">
#40	```bash
#41	git clone https://github.com/mem0ai/mem0.git
#42	cd mem0/examples/docker-compose
#43	```
#44	</Step>
#45	<Step title="Install dependencies for local overrides">
#46	```bash
#47	pip install -r requirements.txt
#48	```
#49	</Step>
#50	</Steps>
#51	</Tab>
#52	</Tabs>
#53
#54	## Define your configuration
#55
#56	<Tabs>
#57	<Tab title="Python">
#58	<Steps>
#59	<Step title="Create a configuration dictionary">
#60	```python
#61	from mem0 import Memory
#62
#63	config = {
#64	"vector_store": {
#65	"provider": "qdrant",
#66	"config": {"host": "localhost", "port": 6333},
#67	},
#68	"llm": {
#69	"provider": "openai",
#70	"config": {"model": "gpt-4.1-mini", "temperature": 0.1},
#71	},
#72	"embedder": {
#73	"provider": "vertexai",
#74	"config": {"model": "textembedding-gecko@003"},
#75	},
#76	"reranker": {
#77	"provider": "cohere",
#78	"config": {"model": "rerank-english-v3.0"},
#79	},
#80	}
#81
#82	memory = Memory.from_config(config)
#83	```
#84	</Step>
#85	<Step title="Store secrets as environment variables">
#86	```bash
#87	export QDRANT_API_KEY="..."
#88	export OPENAI_API_KEY="..."
#89	export COHERE_API_KEY="..."
#90	```
#91	</Step>
#92	</Steps>
#93	</Tab>
#94	<Tab title="config.yaml">
#95	<Steps>
#96	<Step title="Create a `config.yaml` file">
#97	```yaml
#98	vector_store:
#99	provider: qdrant
#100	config:
#101	host: localhost
#102	port: 6333
#103
#104	llm:
#105	provider: azure_openai
#106	config:
#107	api_key: ${AZURE_OPENAI_KEY}
#108	deployment_name: gpt-4.1-mini
#109
#110	embedder:
#111	provider: ollama
#112	config:
#113	model: nomic-embed-text
#114
#115	reranker:
#116	provider: zero_entropy
#117	config:
#118	api_key: ${ZERO_ENTROPY_KEY}
#119	```
#120	</Step>
#121	<Step title="Load the config file at runtime">
#122	```python
#123	from mem0 import Memory
#124
#125	memory = Memory.from_config_file("config.yaml")
#126	```
#127	</Step>
#128	</Steps>
#129	</Tab>
#130	</Tabs>
#131
#132	<Info icon="check">
#133	Run `memory.add(["Remember my favorite cafe in Tokyo."], user_id="alex")` and then `memory.search("favorite cafe", user_id="alex")`. You should see the Qdrant collection populate and the reranker mark the memory as a top hit.
#134	</Info>
#135
#136	## Tune component settings
#137
#138	<AccordionGroup>
#139	<Accordion title="Vector store collections">
#140	Name collections explicitly in production (`collection_name`) to isolate tenants and enable per-tenant retention policies.
#141	</Accordion>
#142	<Accordion title="LLM extraction temperature">
#143	Keep extraction temperatures ≤0.2 so advanced memories stay deterministic. Raise it only when you see missing facts.
#144	</Accordion>
#145	<Accordion title="Reranker depth">
#146	Limit `top_k` to 10–20 results; sending more adds latency without meaningful gains.
#147	</Accordion>
#148	</AccordionGroup>
#149
#150	<Warning>
#151	Mixing managed and self-hosted components? Make sure every outbound provider call happens through a secure network path. Managed rerankers often require outbound internet even if your vector store is on-prem.
#152	</Warning>
#153
#154	## Quick recovery
#155
#156	- Qdrant connection errors → confirm port `6333` is exposed and API key (if set) matches.
#157	- Empty search results → verify the embedder model name; a mismatch causes dimension errors.
#158	- `Unknown reranker` → update the SDK (`pip install --upgrade mem0ai`) to load the latest provider registry.
#159
#160	<CardGroup cols={2}>
#161	<Card
#162	title="Pick Providers"
#163	description="Review the LLM, vector store, embedder, and reranker catalogs."
#164	icon="sitemap"
#165	href="/components/llms/overview"
#166	/>
#167	<Card
#168	title="Deploy with Docker Compose"
#169	description="Follow the end-to-end OSS deployment walkthrough."
#170	icon="server"
#171	href="/cookbooks/companions/local-companion-ollama"
#172	/>
#173	</CardGroup>
#174

z6Mkq5mY3JWtxoxUobWcfNHm7AkRubgSWEZTkBVqZXJviFZ5/my-project-public