# MemLayer

## Docs

- [Get Session Context with GET /memories/context | MemLayer](https://docs.remem.online/api-reference/context.md): GET /memories/context — loads the most relevant memories for a user-agent pair. Use at conversation start to populate the agent's system prompt.
- [Delete a Single Memory — DELETE /memories/{memory_id}](https://docs.remem.online/api-reference/delete.md): DELETE /memories/{memory_id} — permanently removes one memory by ID. Pass user_id and agent_id to verify ownership. Deletion is immediate and irreversible.
- [Wipe All Memories for a User — DELETE /memories | MemLayer](https://docs.remem.online/api-reference/delete-all.md): DELETE /memories — permanently deletes all memories for a user-agent pair. Requires confirm=true. Use for GDPR erasure requests or full agent resets.
- [Duplicate Memory Check — POST /memories/duplicate-check](https://docs.remem.online/api-reference/duplicate-check.md): POST /memories/duplicate-check — checks if a memory already exists without storing it. Returns is_duplicate boolean based on cosine similarity threshold.
- [List All Stored Memories — GET /memories | MemLayer](https://docs.remem.online/api-reference/list.md): GET /memories — paginated listing of every memory stored for a user-agent pair, newest first. Filter by type. Use for audits, dashboards, and data exports.
- [Search Memories Semantically — GET /memories/search](https://docs.remem.online/api-reference/search.md): GET /memories/search — semantic search over stored memories. Returns relevant results ranked by hybrid score. Filter by memory type, top_k, and min_score.
- [Store a New Memory with POST /memories | MemLayer API](https://docs.remem.online/api-reference/store.md): POST /memories — embed and persist a memory for a user-agent pair. Returns the new memory UUID. Near-identical memories are automatically deduplicated.
- [Update a Stored Memory — PATCH /memories/{memory_id}](https://docs.remem.online/api-reference/update.md): PATCH /memories/{memory_id} — updates an existing memory with new content. Re-embeds the content automatically. The memory ID remains unchanged.
- [MemLayer Authentication: API Keys and Request Headers](https://docs.remem.online/authentication.md): MemLayer authenticates every API request via the X-API-Key header. Learn how to obtain your key, store it securely, and handle auth errors.
- [Memory and Request Limits by MemLayer Plan](https://docs.remem.online/billing/limits.md): Understand what counts toward your MemLayer plan limits, how to check your usage, and what happens when you hit a limit.
- [MemLayer Pricing Plans: Free, Pro, and Enterprise](https://docs.remem.online/billing/plans.md): MemLayer offers a free tier to build and test, Pro for production agents, and Enterprise for unlimited scale with your own database.
- [How MemLayer Ranks Memories with Hybrid Scoring](https://docs.remem.online/concepts/hybrid-scoring.md): MemLayer ranks search results with a hybrid score: 70% semantic similarity, 20% recency, and 10% importance — so the best memory always rises to the top.
- [MemLayer Memory Types: Episodic, Semantic, and Summary](https://docs.remem.online/concepts/memory-types.md): MemLayer supports three memory types — episodic, semantic, and summary — each suited for different kinds of agent knowledge and retrieval patterns.
- [Organizing Memories with User IDs and Agent IDs](https://docs.remem.online/concepts/users-and-agents.md): MemLayer scopes every memory to a user_id and agent_id pair. Learn how to structure these identifiers for single-agent and multi-agent architectures.
- [Use MemLayer Asynchronously with AsyncMemLayerClient](https://docs.remem.online/guides/async-usage.md): MemLayer's AsyncMemLayerClient supports fully async operations for FastAPI, LangGraph, and any asyncio application. All methods are awaitable.
- [Handle MemLayer API Errors Gracefully in Your Agent](https://docs.remem.online/guides/error-handling.md): MemLayer raises typed exceptions for authentication failures, plan limits, missing memories, and duplicates — learn how to catch and handle each one.
- [Add Persistent Memory to LangGraph Agents with MemLayer](https://docs.remem.online/guides/langgraph-integration.md): Add persistent memory to any LangGraph graph by wiring MemLayer into load_memory and save_memory nodes that run before and after your agent.
- [Load Session Context at the Start of Every Conversation](https://docs.remem.online/guides/session-context.md): Use MemLayer's context() method to proactively load the most relevant memories at the start of every conversation and inject them into your system prompt.
- [Store, Search, Update and Delete Memories with MemLayer](https://docs.remem.online/guides/store-and-retrieve.md): Learn how to store agent memories, run semantic search, update facts when they change, and delete memories using MemLayer's Python SDK.
- [MemLayer Introduction: AI Agent Memory-as-a-Service](https://docs.remem.online/introduction.md): MemLayer solves the blank-slate problem for AI agents by providing a persistent memory layer that stores, retrieves, and scores memories across sessions.
- [MemLayer Quickstart: Store Your First Memory in 5 Minutes](https://docs.remem.online/quickstart.md): Install the MemLayer Python SDK, store your first memory, and retrieve it with semantic search — all in under five minutes of setup.