What is supermemory™
The universal memory API for the AI era. It helps stop building retrieval from scratch and personalize LLMs for users. It is built for developers who ship.
How to use supermemory™
Supermemory provides a universal memory API. You can add memories by making a POST request to https://api.supermemory.ai/v3/memories with content (text, URL, or PDF) and optional metadata like user_id. Memories can be searched via a GET request to the same endpoint with a query parameter q. You can also connect apps like OneDrive via a POST request to https://api.supermemory.ai/v3/connections/onedrive. Supermemory offers SDKs for Typescript and Python, installable via npm install 'supermemory' or pip install 'supermemory'. You can also integrate by changing your OpenAI base URL to https://api.supermemory.ai/v3/https://api.openai.com/v1/.
Features of supermemory™
- Enterprise-Grade Performance at Any Scale: Built to handle billions of data points with low-latency retrieval, indexing documents, video, or structured product data.
- Seamless Integration Across Teams & Tools: Connects directly to existing stacks like Notion, Google Drive, and custom CRMs with flexible APIs and SDKs.
- Secure by Design. Fully Controllable.: Deployable in the cloud, on-prem, or on-device with full control over data storage.
- Model-agnostic APIs: Works with any LLM provider.
- Sub-400ms latency at scale: Built for speed and scale, re-imagining RAG.
- Best in class performance: Delivers stronger precision and recall at benchmarks.
- Works with AI SDK, Langchain, and more.
- Deploy in a day, not months: SDKs available for Python and Javascript.
Use Cases of supermemory™
- Flow uses supermemory to build the cursor for writing.
- Medtech Vendors uses supermemory to search through 500k vendors.
- Mixus uses Supermemory to power co-intelligence Agentic platform.
