Plugins

Plugins

Plugins extend OpenClaw with new capabilities: channels, model providers, tools, skills, speech, image generation, and more. Some plugins are core (shipped with OpenClaw), others are external (published on npm by the community).

Quick start

1. See what is loaded

   openclaw plugins list
   
2. Install a plugin

   # From npm
   openclaw plugins install @openclaw/voice-call
   
   # From a local directory or archive
   openclaw plugins install ./my-plugin
   openclaw plugins install ./my-plugin.tgz
   
3. Restart the Gateway

   openclaw gateway restart
   

   Then configure under plugins.entries.\<id\>.config in your config file.

If you prefer chat-native control, enable commands.plugins: true and use:

/plugin install clawhub:@openclaw/voice-call
/plugin show voice-call
/plugin enable voice-call

The install path uses the same resolver as the CLI: local path/archive, explicit clawhub:<pkg>, or bare package spec (ClawHub first, then npm fallback).

If config is invalid, install normally fails closed and points you at openclaw doctor --fix. The only recovery exception is a narrow bundled-plugin reinstall path for plugins that opt into openclaw.install.allowInvalidConfigRecovery.

Plugin types

OpenClaw recognizes two plugin formats:

Format	How it works	Examples
Native	openclaw.plugin.json + runtime module; executes in-process	Official plugins, community npm packages
Bundle	Codex/Claude/Cursor-compatible layout; mapped to OpenClaw features	.codex-plugin/, .claude-plugin/, .cursor-plugin/

Both show up under openclaw plugins list. See Plugin Bundles for bundle details.

If you are writing a native plugin, start with Building Plugins and the Plugin SDK Overview.

Official plugins

Installable (npm)

Plugin	Package	Docs
Matrix	@openclaw/matrix	Matrix
Microsoft Teams	@openclaw/msteams	Microsoft Teams
Nostr	@openclaw/nostr	Nostr
Voice Call	@openclaw/voice-call	Voice Call
Zalo	@openclaw/zalo	Zalo
Zalo Personal	@openclaw/zalouser	Zalo Personal

Core (shipped with OpenClaw)

Model providers (enabled by default)

anthropic, byteplus, cloudflare-ai-gateway, github-copilot, google, huggingface, kilocode, kimi-coding, minimax, mistral, modelstudio, moonshot, nvidia, openai, opencode, opencode-go, openrouter, qianfan, synthetic, together, venice, vercel-ai-gateway, volcengine, xiaomi, zai

Memory plugins

• memory-core — bundled memory search (default via plugins.slots.memory)
• memory-lancedb — install-on-demand long-term memory with auto-recall/capture (set plugins.slots.memory = "memory-lancedb")

Speech providers (enabled by default)

elevenlabs, microsoft

Other

• browser — bundled browser plugin for the browser tool, openclaw browser CLI, browser.request gateway method, browser runtime, and default browser control service (enabled by default; disable before replacing it)
• copilot-proxy — VS Code Copilot Proxy bridge (disabled by default)

Looking for third-party plugins? See Community Plugins.

Configuration

{
  plugins: {
    enabled: true,
    allow: ["voice-call"],
    deny: ["untrusted-plugin"],
    load: { paths: ["~/Projects/oss/voice-call-extension"] },
    entries: {
      "voice-call": { enabled: true, config: { provider: "twilio" } },
    },
  },
}

Field	Description
enabled	Master toggle (default: true)
allow	Plugin allowlist (optional)
deny	Plugin denylist (optional; deny wins)
load.paths	Extra plugin files/directories
slots	Exclusive slot selectors (e.g. memory, contextEngine)
entries.\<id\>	Per-plugin toggles + config

Config changes require a gateway restart. If the Gateway is running with config watch + in-process restart enabled (the default openclaw gateway path), that restart is usually performed automatically a moment after the config write lands.

Plugin states: disabled vs missing vs invalid

• Disabled: plugin exists but enablement rules turned it off. Config is preserved.
• Missing: config references a plugin id that discovery did not find.
• Invalid: plugin exists but its config does not match the declared schema.

Discovery and precedence

OpenClaw scans for plugins in this order (first match wins):

1. Config paths

   plugins.load.paths — explicit file or directory paths.

2. Workspace extensions

   \<workspace\>/.openclaw/<plugin-root>/*.ts and \<workspace\>/.openclaw/<plugin-root>/*/index.ts.

3. Global extensions

   ~/.openclaw/<plugin-root>/*.ts and ~/.openclaw/<plugin-root>/*/index.ts.

4. Bundled plugins

   Shipped with OpenClaw. Many are enabled by default (model providers, speech). Others require explicit enablement.

Enablement rules

• plugins.enabled: false disables all plugins
• plugins.deny always wins over allow
• plugins.entries.\<id\>.enabled: false disables that plugin
• Workspace-origin plugins are disabled by default (must be explicitly enabled)
• Bundled plugins follow the built-in default-on set unless overridden
• Exclusive slots can force-enable the selected plugin for that slot

Plugin slots (exclusive categories)

Some categories are exclusive (only one active at a time):

{
  plugins: {
    slots: {
      memory: "memory-core", // or "none" to disable
      contextEngine: "legacy", // or a plugin id
    },
  },
}

Slot	What it controls	Default
memory	Active memory plugin	memory-core
contextEngine	Active context engine	legacy (built-in)

CLI reference

openclaw plugins list                       # compact inventory
openclaw plugins list --enabled            # only loaded plugins
openclaw plugins list --verbose            # per-plugin detail lines
openclaw plugins list --json               # machine-readable inventory
openclaw plugins inspect <id>              # deep detail
openclaw plugins inspect <id> --json       # machine-readable
openclaw plugins inspect --all             # fleet-wide table
openclaw plugins info <id>                 # inspect alias
openclaw plugins doctor                    # diagnostics

openclaw plugins install <package>         # install (ClawHub first, then npm)
openclaw plugins install clawhub:<pkg>     # install from ClawHub only
openclaw plugins install <spec> --force    # overwrite existing install
openclaw plugins install <path>            # install from local path
openclaw plugins install -l <path>         # link (no copy) for dev
openclaw plugins install <plugin> --marketplace <source>
openclaw plugins install <plugin> --marketplace https://github.com/<owner>/<repo>
openclaw plugins install <spec> --pin      # record exact resolved npm spec
openclaw plugins install <spec> --dangerously-force-unsafe-install
openclaw plugins update <id>             # update one plugin
openclaw plugins update <id> --dangerously-force-unsafe-install
openclaw plugins update --all            # update all
openclaw plugins uninstall <id>          # remove config/install records
openclaw plugins uninstall <id> --keep-files
openclaw plugins marketplace list <source>
openclaw plugins marketplace list <source> --json

openclaw plugins enable <id>
openclaw plugins disable <id>

Bundled plugins ship with OpenClaw. Many are enabled by default (for example bundled model providers, bundled speech providers, and the bundled browser plugin). Other bundled plugins still need openclaw plugins enable <id>.

--force overwrites an existing installed plugin or hook pack in place. It is not supported with --link, which reuses the source path instead of copying over a managed install target.

--pin is npm-only. It is not supported with --marketplace, because marketplace installs persist marketplace source metadata instead of an npm spec.

--dangerously-force-unsafe-install is a break-glass override for false positives from the built-in dangerous-code scanner. It allows plugin installs and plugin updates to continue past built-in critical findings, but it still does not bypass plugin before_install policy blocks or scan-failure blocking.

This CLI flag applies to plugin install/update flows only. Gateway-backed skill dependency installs use the matching dangerouslyForceUnsafeInstall request override instead, while openclaw skills install remains the separate ClawHub skill download/install flow.

Compatible bundles participate in the same plugin list/inspect/enable/disable flow. Current runtime support includes bundle skills, Claude command-skills, Claude settings.json defaults, Claude .lsp.json and manifest-declared lspServers defaults, Cursor command-skills, and compatible Codex hook directories.

openclaw plugins inspect <id> also reports detected bundle capabilities plus supported or unsupported MCP and LSP server entries for bundle-backed plugins.

Marketplace sources can be a Claude known-marketplace name from ~/.claude/plugins/known_marketplaces.json, a local marketplace root or marketplace.json path, a GitHub shorthand like owner/repo, a GitHub repo URL, or a git URL. For remote marketplaces, plugin entries must stay inside the cloned marketplace repo and use relative path sources only.

See openclaw plugins CLI reference for full details.

Plugin API overview

Native plugins export an entry object that exposes register(api). Older plugins may still use activate(api) as a legacy alias, but new plugins should use register.

export default definePluginEntry({
  id: "my-plugin",
  name: "My Plugin",
  register(api) {
    api.registerProvider({
      /* ... */
    });
    api.registerTool({
      /* ... */
    });
    api.registerChannel({
      /* ... */
    });
  },
});

OpenClaw loads the entry object and calls register(api) during plugin activation. The loader still falls back to activate(api) for older plugins, but bundled plugins and new external plugins should treat register as the public contract.

Common registration methods:

Method	What it registers
registerProvider	Model provider (LLM)
registerChannel	Chat channel
registerTool	Agent tool
registerHook / on(...)	Lifecycle hooks
registerSpeechProvider	Text-to-speech / STT
registerMediaUnderstandingProvider	Image/audio analysis
registerImageGenerationProvider	Image generation
registerWebSearchProvider	Web search
registerHttpRoute	HTTP endpoint
registerCommand / registerCli	CLI commands
registerContextEngine	Context engine
registerService	Background service

Hook guard behavior for typed lifecycle hooks:

• before_tool_call: { block: true } is terminal; lower-priority handlers are skipped.
• before_tool_call: { block: false } is a no-op and does not clear an earlier block.
• before_install: { block: true } is terminal; lower-priority handlers are skipped.
• before_install: { block: false } is a no-op and does not clear an earlier block.
• message_sending: { cancel: true } is terminal; lower-priority handlers are skipped.
• message_sending: { cancel: false } is a no-op and does not clear an earlier cancel.

For full typed hook behavior, see SDK Overview.

Related

• Building Plugins — create your own plugin
• Plugin Bundles — Codex/Claude/Cursor bundle compatibility
• Plugin Manifest — manifest schema
• Registering Tools — add agent tools in a plugin
• Plugin Internals — capability model and load pipeline
• Community Plugins — third-party listings