Supermemory
Memory engine for AI agents that turns conversations, files and URLs into facts and injects them into a model's context.
Supermemory is a memory layer for AI agents. Developers send Supermemory conversations, files, URLs or raw data; its own model extracts facts, keeps them current in a vector-graph database and injects relevant memories into the agent's context. Supermemory is used through an API, plugins or MCP, with usage billed against monthly credits.
Best for Developers building agents or AI apps that need long-term, per-user memory that works across different models.
- Platforms
- API, MCP, Plugin
- Pricing model
- Freemium
- Free plan
- Yes
- API
- Yes
- Made by
- Supermemory
- Added
Last verified: . Information taken from supermemory.ai.
What is Supermemory?
Supermemory is infrastructure that gives AI agents persistent memory. A developer hands it whatever the agent sees, such as chats, files, web addresses or the contents of a drive. The company's model, learner-1, converts that material into facts, links them to one another and to a timeline, and revises or discards them as new information arrives. Stored memories live in a vector-graph database and are returned as a profile, a search result or a file, then placed into the model's context while it works. Since the memory sits outside any model provider, the same store can serve several models and agent harnesses. Supermemory is reached through an API, plugins and MCP, and can be deployed in a customer's data centre, VPC or laptop. Plans are metered by usage with a monthly credit balance.
Key Features
Memory Graph per User
Builds profiles and hierarchies of facts for each user that agents can draw on during a session.
learner-1 Model
Extracts facts from incoming content and keeps the memory store current by revising, combining and removing entries.
SuperRAG
Handles multimodal extraction, chunking and retrieval without the developer managing embeddings, and is also offered as a filesystem.
Search and Traversal
Runs semantic search and graph traversal over stored content in a single call.
Connectors
Adds sources such as Gmail on the Max plan, and S3 and a web crawler on Scale.
Flexible Deployment
Runs as a managed service or self-hosted, including air-gapped setups on the Enterprise tier.
Screenshots
Use Cases
- Personalised Assistants
- An app remembers each user's preferences and history between sessions.
- Coding Agents
- A coding agent keeps project knowledge without re-reading every file on each task.
- Support Bots
- A customer-facing agent recalls earlier conversations with the same person.
- Document-Heavy Retrieval
- A team indexes PDFs, images, audio and video for an agent to query.
- Regulated Environments
- An organisation runs the memory system inside its own network.
Things to know
- Billed in SM tokens
- Charges are based on the unique content ingested; for supported updates to the same document, unchanged content is not charged again.
- Credits are required to keep running
- Usage needs an available balance. More credits can be bought in the console, and paid plans can top up automatically within a set limit.
- PDF extraction differs by plan
- Page-by-page OCR with figure and diagram descriptions is included on Scale and Enterprise; other plans use standard extraction.
- Startup and research offer
- Qualifying early-stage startups and academic teams can apply for three months of the Scale plan at no cost.
FAQs
What is Supermemory?
Supermemory is a memory engine for AI agents. It stores facts extracted from your data and supplies them to a model's context when they are relevant.
What is Supermemory used for?
Supermemory is used to give agents and AI applications long-term memory of users, tasks and documents, and to run retrieval over stored content.
Is Supermemory free?
Supermemory: a free plan is available. Paid plan rates are published on supermemory.ai. Check the official pricing page for current amounts.
How much does Supermemory cost?
Supermemory: a free plan is available. Paid plan rates are published on supermemory.ai. Check the official pricing page for current amounts.
Does Supermemory have an API?
Yes. Supermemory is available through an API, as well as through plugins and MCP.
Can Supermemory be self-hosted?
Yes. Supermemory offers self-hosted deployments on the Scale and Enterprise plans, and Enterprise also covers fully air-gapped installations.
Do unused Supermemory credits roll over?
No for subscription credits, which reset every month. Credits bought as a top-up on Supermemory do not expire.
Is Supermemory compliant with security standards?
Supermemory reports SOC 2 certification, compliance with HIPAA and alignment with GDPR, and says it signs customer contracts and data processing agreements.
Who it's for
- AI developers
- Agent builders
- Startups
- Enterprise engineering teams
Comments
No comments yet. If you have used Supermemory, tell other readers how it went.
Alternatives to Supermemory
| Tool | Free plan | Starting price | API |
|---|---|---|---|
 ZepContext and memory layer for AI agents that builds governed knowledge graphs from business data, documents and conversations. |
Yes | — | Yes |
 DocsBotPlatform for building AI agents and chatbots trained on your documentation to answer questions and take actions. |
Yes | — | Yes |
 n8nWorkflow automation platform for building AI agents and automations visually, with code steps and self-hosting. |
— | — | — |
 FirecrawlWeb data API that searches, scrapes, crawls and interacts with websites and returns clean output for AI applications. |
Yes | $16/mo | Yes |
 Kapa.aiIngestion and retrieval platform that turns company docs, tickets and wikis into AI agents and a search API. |
No | — | Yes |
Listed in the same categories as Supermemory. A starting price appears only where we checked it on the maker’s pricing page; a dash means not checked or not published. All Developer API tools