pumaDB
A lightweight, database-free memory API layer for AI agents.
āļ āļēāļāļĢāļ§āļĄ
pumaDB āļāļ·āļāļāļ°āđāļĢ?
pumaDB is a lightweight, hosted JSON memory layer and API designed specifically for AI agents. It provides a simple, shared place for agents to save and reuse notes, facts, user preferences, project context, transcripts, and task states across different sessions, tools, and chats. It eliminates the need for managing complex infrastructure like databases, vector stores, or custom RAG stacks, offering both a hosted Model Context Protocol (MCP) and a server-side REST API.
pumaDB āđāļāđāļŠāļģāļŦāļĢāļąāļāļāļ°āđāļĢ?
āđāļāđāļ

āļāļļāļāļŠāļĄāļāļąāļāļīāļŦāļĨāļąāļ
- Durable JSON memory without database setup or infrastructure management
- Dual connectivity options via hosted MCP or a server-side REST API
- Automatic version history saving the last 10 versions for 30 days
- Natural language edits that translate plain text requests into filtered updates
- Built-in safety rails including scoped table limits and per-key rate limits
āļ§āļīāļāļĩāđāļāđ pumaDB?
To use pumaDB, you can either connect via a hosted Model Context Protocol (MCP) endpoint by adding it to clients like Codex, ChatGPT, or Claude, or you can interact directly with its server-side REST API by utilizing a secure backend key to perform JSON row operations like remember, query, and update.
āđāļāļĢāļ·āđāļāļāļĄāļ·āļāļāļēāļāđāļĨāļ·āļāļ
āļāđāļāļāļĩ & āļāđāļāđāļŠāļĩāļĒ
āđāļĄāđāļĄāļĩāļāđāļāļĄāļđāļĨ
āļāļĢāļāļĩāļāļēāļĢāđāļāđāļāļēāļ
- Storing user preferences across separate AI agent chat sessions
- Maintaining reusable operating instructions and project conventions for coding agents
- Tracking task states, open threads, and handoff notes for long-running workflows
- Saving research clippings, summaries, and links from continuous investigations
āļĢāļĩāļ§āļīāļ§āļāļđāđāđāļāđ
āļĒāļąāļāđāļĄāđāļĄāļĩāļĢāļĩāļ§āļīāļ§
āđāļāđāļāļāļāđāļĢāļāļāļĩāđāļĢāļĩāļ§āļīāļ§āđāļāļĢāļ·āđāļāļāļĄāļ·āļāļāļĩāđ!