
by GibsonAI
Memori is a memory layer for AI agents that stores and recalls facts from conversations and agent actions. It ships Python and TypeScript SDKs, a cloud service and an MCP endpoint.
APIFree
Step 1: Install the SDK with pip install memori or npm install @memorilabs/memori.
Step 2: Sign up at app.memorilabs.ai and set MEMORI_API_KEY alongside your LLM provider key.
Step 3: Register your LLM client with Memori and set attribution with an entity ID and a process ID.
Step 4: Make LLM calls as usual; conversations are saved and recalled automatically.
Step 5: Check memories and usage in the dashboard, or run python -m memori quota to see your quota.
Step 6: For coding agents, connect through MCP, for example with claude mcp add using the Memori endpoint and your API key.
Memori ties memories to an entity and a process. Without attribution it cannot create memories.
Per the README, Advanced Augmentation is always free for developers and works without an account at limited rates. An API key gives higher limits.
Yes. The README points to a bring-your-own-database option described in its docs.
Yes. Claude Code, Cursor, Codex, Warp and Antigravity can connect through the MCP endpoint without an SDK integration.
OraichainOraichain is an AI-focused layer 1 blockchain and oracle that lets smart contracts use external AI APIs and models. It also hosts an ecosystem of AI and DeFi products.