
AI Agent for Membase
Integrate FlowHunt’s AI Agent with Membase, the memory layer for AI agents. Membase keeps one long-term memory that your agents and AI tools (Claude Code, Cursor, Codex, ChatGPT) search and update, so context carries across sessions and tools.

Shared Long-Term Memory
Add notes, documents, web pages and chat exports once; Membase learns them into memories organized by topic. Every search result shows the source it came from.
- One Memory for Every AI:
- Your FlowHunt agents and AI tools like Claude Code, Cursor, Codex and ChatGPT read from and write to the same memory.
- Scoped Access per Agent:
- Give each agent its own key with an access level and the set of memories it can reach.
- Traceable Answers:
- Every search result shows the source it came from.
- Hosted or Local:
- Use the hosted service or run Membase locally.

Flexible Integration & Control
Connect over MCP at https://api.app.membase.com/mcp-http with an API key, or use the Python and TypeScript SDKs and REST API. Each key has an access level and a set of memories it can reach.
- MCP over HTTP:
- Connect any MCP-compatible agent to Membase with a single endpoint and an API key.
- SDKs and REST API:
- Build on Membase with the Python and TypeScript SDKs or call the REST API directly.
- Per-Key Access Control:
- Each API key has an access level and a defined set of memories it can reach.
MCP INTEGRATION
Available Membase MCP Integration Tools
The following tools are available as part of the Membase MCP integration:
- search_memories
Search across the memories a key can reach.
- add_memory
Save a fact or note.
- add_document
Add a document for Membase to learn from.
- list_containers
Browse the memories a key can reach.
- list_documents
Browse the documents Membase has learned from.
- get_profile
Read the user’s standing preferences.
- forget_memory
Remove a memory (full-access keys, with confirmation).
- delete_document
Remove a document (full-access keys, with confirmation).
Connect Your Membase with FlowHunt AI
Give your FlowHunt agents a long-term memory they share with the rest of your AI tools.

What is Membase
Membase, built by Unibase, is the memory layer for AI agents. It gives agents and AI tools one shared long-term memory, reachable over MCP, REST, or Python and TypeScript SDKs. Free during beta. Learn more at https://membase.com.
Capabilities
What we can do with Membase
Membase turns your conversations, files and connected apps into one long-term memory that you control, and lets you decide which AI tools and agents can use it.
- Add Knowledge from Anywhere
- Save AI conversations and web pages with the Chrome extension, upload PDFs and notes, connect apps, or let your own agents add memories over MCP or the API.
- Organized by Topic
- Membase learns what you add into memories organized by topic, such as project decisions, reading notes or customers.
- Review and Correct
- See what Membase remembers, fix mistakes and remove outdated information at any time.
- Control Who Sees What
- Choose which AI can use which memories and keep personal, project and client knowledge separate.
- Hosted or Local
- Use the hosted service, or run Membase locally with the same API and MCP tools when data must stay on your machine.

How AI Agents Benefit from Membase
Without memory, every agent run starts from zero. With Membase, your FlowHunt agents can save findings and retrieve relevant knowledge as work progresses, recall user preferences and past decisions, and share selected memories with other agents and AI tools such as Claude Code, Cursor, Codex and ChatGPT. Scoped API keys keep each agent limited to the memories it needs, and every search result points back to its source.