New · AI memory management

Your AI agents' memory,

HippoKairn centralizes, secures and versions your AI agents and all their tooling (Skills, MCP servers, hooks and AI Core). Compatible with every AI tool on the market.

CLAUDE.md.cursorrulesGEMINI.mdAGENTS.mdHooksMCP Servers*.json · *.yml
0+
Devs
0+
Mémoires
< 0 min
Setup
Works with your AI tools
Claude CodeCursorWindsurfGitHub CopilotGemini CLIAiderClineContinueCodexAmpZedClaude CodeCursorWindsurfGitHub CopilotGemini CLIAiderClineContinueCodexAmpZed
Agents

Your agents, versioned and ready to ship

Each agent bundles its model, allowed tools and tooling (Skills, MCP, hooks). Manage them like code.

Backend Reviewer

Backend code review

v2.1
Claude Opus 4.6
Tools
ReadGrepBash
3 Skills2 MCP1 Hooks

Docs Writer

Documentation writing

v1.4
Claude Sonnet 4.6
Tools
ReadWrite
2 Skills1 MCP0 Hooks

PR Triage

Pull request triage

v3.0
Claude Haiku 4.5
Tools
ReadGrep
1 Skills3 MCP2 Hooks

Security Auditor

Security audit

v1.2
Claude Opus 4.6
Tools
ReadGrepBash
4 Skills1 MCP2 Hooks
Features

Everything you need

A complete platform to manage your AI agents' memory with confidence.

Enhanced security

Protection against prompt injection, access control and fine-grained memory resource management at enterprise scale.

Simple integration

Connect your AI agents in a few clicks. Compatible with your config files, hooks and MCP servers with no complex setup.

All your agents in one place

One hub for your agents and all their tooling: Skills, MCP servers, hooks and AI Core configuration (CLAUDE.md), organized and accessible to the whole team.

Smart versioning

Full history of all changes. Compare, restore and continuously improve your AI agents.

CLAUDE.md

Custom instructions, roles, project context

# Agent Instructions
## Role: Senior Dev
- Always write tests
memory.md

Persistent memories, agent facts and preferences

## User Preferences
- Prefers TypeScript
- No semicolons
Hooks

Automatic actions before/after each interaction

PreToolUse → validate
PostToolUse → log
Stop → notify
MCP Servers

Connections to external tools and data sources

filesystem: enabled
github: connected
postgres: ready
AGENTS.mduniversal format

HippoKairn handles any .md, .json or .yml file. Your AI tool isn't listed? Doesn't matter, it works anyway.

Claude CodeCursorWindsurfCopilotClineAmpCodexGemini CLIAiderContinueZed
What makes us different

Why HippoKairn isn't just another Notion folder

Four principles that mean we treat AI memory like code, not like docs.

🧬
diff #00101 / 04

Memory-as-Code

Your agents, their Skills, MCPs, hooks and AI Core are versioned like your repo. Diff, rollback, branches: same rigor, same workflow.

🔐
diff #00202 / 04

Secure by default

Prompt injection detection on write, fine-grained MCP scoping, full audit log. Compliance-ready from day one.

🤝
diff #00303 / 04

Built for teams

Group sharing, hook review before deploy, per-project profiles. No more CLAUDE.md drift across the team.

🌐
diff #00404 / 04

Universal multi-agent

One format, every tool. Claude, Cursor, Windsurf, Copilot, Gemini, Aider, even the ones that don't exist yet.

How it works

Simple to set up

Up and running in under 5 minutes.

STEP 1

Connect your workspace

Link HippoKairn to your AI tools in just a few clicks.

STEP 2

Compose your agents

Create your agents and orchestrate their Skills, MCP servers, hooks and AI Core configs (CLAUDE.md) from a unified interface.

STEP 3

Collaborate as a team

Share memories, manage access by groups and track changes in real time.

FAQ

Frequently asked questions

Everything you need to know before getting started.

Still have questions?

hello@hippokairn.fr

Ready to master your AI agents' memory?

Join the teams that trust HippoKairn to secure and optimize their AI agents.