Memorix is a shared project memory layer. Agents connect through the interfaces they already support: plugin packages, MCP, project instructions, hooks, skills, or the bundled memcode terminal agent.

For most users, start with:

memorix setup --agent <agent> --global

That command installs the recommended Memorix package or config for the target agent. Manual integrate, hooks, and raw MCP config remain available when you need a fallback setup.

HTTP is not required for normal agent setup. Use it when you intentionally want a shared background process, dashboard, VPS Docker deployment, or multiple clients using the same MCP endpoint. Local verification runs the Node service directly; Docker is not required on a developer machine.

CLI, MCP, and HTTP have separate jobs:

• CLI is the direct command surface. Use it for setup, diagnostics, memory operations, Git Memory, import/export, dashboard commands, and orchestration.
• Stdio MCP is the normal agent/IDE bridge. It gives the agent live Memorix tools by launching memorix serve.
• HTTP MCP is the shared service bridge. Use it for one endpoint shared by multiple clients, dashboard, Docker, or supervised foreground debugging.

Generated guidance also has scope:

• Project guidance (CLAUDE.md, AGENTS.md, GEMINI.md, Cursor/Windsurf/Kiro/Trae rules, DeepSeek Harness AGENTS.md) may say the repository is configured for Memorix.
• Plugin or skill package guidance is workspace-safe. It says to use Memorix when the active workspace has Memorix tools available.
• --global writes user-level plugin, config, and hook surfaces where the host supports them.
• Running the same setup command without --global is the repo-local path for one project only.
• Hooks are only global where the host provides a global hook surface. If the host only supports project hooks, Memorix uses that project scope instead.

Install Memorix:

npm install -g memorix
memorix init --global

Then install the agent integration you use:

memorix setup --agent claude --global
memorix setup --agent codex --global
memorix setup --agent copilot --global
memorix setup --agent cursor --global
memorix setup --agent pi --global
memorix setup --agent gemini-cli --global
memorix setup --agent opencode --global
memorix setup --agent windsurf --global
memorix setup --agent kiro --global
memorix setup --agent antigravity --global
memorix setup --agent trae --global
memorix setup --agent openclaw --global
memorix setup --agent hermes --global
memorix setup --agent codebuddy --global
memorix setup --agent omp --global
memorix setup --agent dsh --global

Use memorix setup --agent all --global only when you intentionally want every supported agent integration generated for the current machine/user scope.

If an agent only needs a raw stdio MCP entry:

{
  "mcpServers": {
    "memorix": {
      "command": "memorix",
      "args": ["serve"]
    }
  }
}

Avoid npx in persistent MCP configs. Use the globally installed memorix binary so startup is predictable.

For npm-based MCP clients or registry testers, the complete package command is npx -y memorix serve. Do not omit serve: bare memorix is the interactive memcode entry in a human terminal.

For HTTP mode:

memorix background start

Endpoint:

http://localhost:3211/mcp

In HTTP mode, bind each project session with memorix_session_start(projectRoot=...) when the client can provide the workspace path.

These commands remain useful for controlled fallback setups:

memorix integrate --agent cursor
memorix integrate --agent gemini-cli
memorix setup --agent openclaw --global
memorix setup --agent hermes --global
memorix setup --agent omp --global
memorix hooks install --agent cursor
memorix hooks install --agent opencode

Use them when you do not want the full setup package, or when you are updating one generated integration file by hand.

memorix integrate --agent <agent> writes usage guidance and MCP settings where supported. memorix setup --agent openclaw|hermes|omp installs the official package or plugin entry that those hosts expose, plus MCP config where needed. memorix hooks install --agent <agent> installs fallback automatic capture only where the agent exposes standalone hook files; package-owned hooks stay managed by memorix setup.

Shared files such as AGENTS.md, CLAUDE.md, and GEMINI.md are appended or updated carefully so existing project instructions are not replaced wholesale.

Memorix ships an official skill set for agents that support skills:

• memorix-memory
• memorix-reasoning
• memorix-sessions
• memorix-git-memory
• memorix-mini-skills
• memorix-orchestrate
• memorix-troubleshooting

These skills are operational guidance for agents: when to search, when to store, when to use CLI fallbacks, when Git Memory is evidence, and when orchestration coordination is appropriate.

Memorix can also promote durable observations into reusable mini-skills. Use this when a project pattern, gotcha, or workflow should become guidance that agents can rediscover later.

Useful commands and tools:

memorix skills

MCP tools:

• memorix_skills
• memorix_promote
• memorix_rules_sync

Plugin and package integrations include the official Memorix skill set where the agent supports skills.

Memorix does three things:

1. Stores project memory locally and makes it searchable.
2. Exposes that memory through plugin packages, MCP, CLI, SDK, hooks, rules, and skills.
3. Provides memcode as a bundled terminal agent for users who want a Memorix-powered coding session in the terminal.

Different agents expose different extension points. When an agent supports plugins, Memorix uses plugins. When it uses rules or instruction files, Memorix writes those. When it only speaks MCP, MCP is the integration.