GPU Brain
gpubrain.app.nz- Category
- AI Agents
- Rank
- No. 1385Tools index
Previous survey · No. 1297 ·
- Pricing
- Freemium
- Type
- APP
- Builder
- @LeeLeepenkman
- Date
About
A hosted memory layer that gives AI agents persistent, searchable memory across conversations. Agents can search, read, write, and link information in markdown format through MCP, CLI, and API access.
What it does
The site pitches a hosted memory service that agents query before responding, then update afterward with new facts, links, and tags. Humans can browse and edit the same markdown pages through a web interface, keeping the stored knowledge auditable by people, not only machines.
Stated on the product site
- Interface
- Offers a hosted MCP server alongside a local MCP option, a CLI, and a REST API, all pointed at the same underlying memory store.
- Pricing
- Signup includes a free credit allotment, with a paid tier billed monthly or annually.
- Data model
- Memory is organized as markdown pages that can be typed and connected into a graph via links and backlinks, rather than kept as flat notes.
- Integrations
- Positions itself as compatible with multiple MCP clients including Claude and Cursor, connected through a config snippet naming an API key and CLI command.
- Search
- Combines several retrieval methods, described as static embeddings alongside keyword and tag matching, rather than relying on one search method alone.
Not stated on the site
- The page does not show how memory is stored or hosted beyond calling it a hosted layer.
- Data retention, privacy, and export policies are not described on the page.
- The full pricing structure beyond the Pro tier price and free credits is not detailed on the page.
- The size or performance limits of the memory store (how much can be written or searched) are not stated.
Written from the product site at gpubrain.app.nz.
What it can do
Store conversation data persistently
Conversation history and context data → Persistent memory storage for AI agents
Search through stored memory
Search query or keywords → Relevant stored information and context
Write information to memory
New data or information in markdown format → Stored memory entry
Read stored information
Memory location or identifier → Retrieved information in markdown format
Link related information together
Multiple pieces of stored information → Connected memory nodes with relationships
Access memory through API calls
API requests with authentication → JSON responses with memory data
Manage memory through command line interface
CLI commands and parameters → Command execution results and memory operations
Tags
Media
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.