MCP server

other

pumaDB gives Neotask's agent a durable, hosted place to put data that needs to survive between runs — a memory layer built for AI agents or small server-side apps. Store and retrieve JSON rows through its REST API or MCP server, so an agent or app can create, read, and manage persistent records instead of losing state between sessions.

• Store and retrieve durable JSON rows as a hosted memory layer
• Built for AI agents or small server side apps that need persistent state
• Create, read, and manage records through a REST API or MCP server

What you can automate

JSON record storage

pumaDB lets you store and retrieve durable JSON rows as a hosted memory layer for AI agents or small server side apps. You can create, read, and manage persistent records through its REST API or MCP server.

Real workflows

Give an agent memory that survives between sessions

Ask Neotask to save a set of preferences or task state as JSON records in pumaDB; the next session it can retrieve those same records instead of starting from scratch.

Frequently asked questions

What does the pumaDB connection actually do?

pumaDB lets you store and retrieve durable JSON rows as a hosted memory layer for AI agents or small server side apps. You can create, read, and manage persistent records through its REST API or MCP server.

Is this a full database or just simple storage?

It's built specifically for durable JSON row storage and retrieval — a lightweight persistent memory layer, not a full relational database.