OpenClaw ships with over 90 plugins spanning LLM providers, messaging channels, search tools, media generation, memory systems, and infrastructure. Here's every one of them, categorized.
If you are weighing OpenClaw against the alternatives, our comparison database keeps a live scorecard of how the leading AI Agent tools stack up.
TL;DR: OpenClaw has plugins for ~any model provider you can name, ~every messaging platform that matters, search tools for days, media generation pipelines, speech TTS/ASR, memory backends, and infrastructure glue. If you're looking at OpenClaw, this is the full map.
How to Use Plugins
Plugins in OpenClaw are configured through the openclaw.yaml configuration file. Each plugin can be enabled or disabled with a simple toggle. Most plugins require minimal configuration — an API key for external services or a channel token for messaging platforms. For example, enabling the Telegram channel plugin requires just a bot token from @BotFather, while enabling a search plugin like Brave Search needs a free API key from the Brave Search API dashboard.
When a plugin is enabled, OpenClaw automatically discovers its capabilities and makes them available to the AI agent. The agent can then use the plugin's functionality autonomously — searching the web, generating images, reading files, or sending messages — as part of its reasoning loop. You don't need to write code to integrate plugins; configuration is all that's required.
The plugin ecosystem is one of OpenClaw's strongest differentiators. With over 100 plugins available, it covers more integrations than most competing open-source agent platforms. The community actively contributes new plugins through the ClawHub marketplace, and the plugin API is well-documented for developers who want to build custom integrations.
Messaging Channels
OpenClaw connects to your messaging apps. These channel plugins handle message sending, receiving, and interactive UI (buttons, markdown, etc.).
| Plugin | Platform |
|---|---|
| telegram | Telegram (bots + user accounts) |
| discord | Discord bots |
| signal | Signal messenger |
| slack | Slack |
| msteams | Microsoft Teams |
| matrix | Matrix protocol |
| irc | IRC (old-school) |
| nostr | Nostr (NIP-04 encrypted DMs) |
| imessage | Apple iMessage |
| bluebubbles | BlueBubbles (iMessage bridge) |
| line | LINE messenger |
| feishu | Feishu/Lark (by @m1heng) |
| qqbot | QQ Bot |
| googlechat | Google Chat |
| mattermost | Mattermost |
| zalo | Zalo messenger |
| zalouser | Zalo Personal Account (via zca-js) |
| xmpp | XMPP chat |
| twitch | Twitch chat |
| tlon | Tlon / Urbit |
| nextcloud-talk | Nextcloud Talk |
| synology-chat | Synology Chat |
| webhooks | Generic webhook bridge |
Total: 24 channel plugins
Speech & Voice
Text-to-speech, speech-to-text, and voice call plugins.
| Plugin | What It Does |
|---|---|
| elevenlabs | ElevenLabs TTS (high-quality voice synthesis) |
| deepgram | Deepgram ASR/TTS (media understanding) |
| microsoft | Microsoft Speech (Azure TTS) |
| gradium | Gradium speech |
| speech-core | Speech runtime runtime core |
| voice-call | Voice call plugin |
| talk-voice | Talk voice support |
Total: 7 speech plugins
Memory & Knowledge
Persistent memory, wikis, and knowledge management.
| Plugin | What It Does |
|---|---|
| memory-core | Core memory search plugin |
| memory-lancedb | LanceDB-backed long-term memory with auto-recall/capture |
| memory-wiki | Persistent wiki vault (Obsidian-compatible) |
| active-memory | Active memory management |
Total: 4 memory plugins
Infrastructure & Runtime
Plugins for managing the host, sandbox, discovery, and observability.
| Plugin | What It Does |
|---|---|
| openshell | OpenShell sandbox backend (exec isolation) |
| browser | Browser automation tool (Playwright-based) |
| bonjour | mDNS/gateway discovery |
| diagnostics-otel | OpenTelemetry exporter |
| device-pair | Device pairing |
| thread-ownership | Thread ownership management |
| skill-workshop | Skill workshop (build/audit skills) |
| qa-lab | QA lab |
| qa-channel | QA channel |
| phone-control | Phone control |
| copilot-proxy | Copilot proxy |
Total: 11 infrastructure plugins
By the Numbers
| Category | Count |
|---|---|
| Model Providers | 42 |
| Messaging Channels | 24 |
| Search & Web | 10 |
| Development Tools | 10 |
| Infrastructure | 11 |
| Speech & Voice | 7 |
| Media Generation | 7 |
| Memory & Knowledge | 4 |
| Communication Bridges | 3 |
| Total | ~118 |
*Last updated: May 2026. OpenClaw v0.x. Versions and availability may change. Check the official docs for the latest list.*
The OpenClaw plugin ecosystem continues to grow rapidly, with community contributions arriving through the ClawHub marketplace. New plugins are added weekly, covering emerging AI providers, messaging platforms, and infrastructure tools. If you need a plugin that doesn't exist yet, the plugin API is well-documented and several community members have published tutorials for building custom integrations.
Plugin development follows a standardized pattern: a manifest file declares the plugin's capabilities, a handler module implements the logic, and configuration options are exposed through YAML. Most plugins are 200-500 lines of Python, making them accessible to intermediate developers who want to contribute. The OpenClaw maintainers review all submissions to ensure quality and security before they appear in the marketplace.
For teams managing multiple OpenClaw instances, the plugin system supports centralized configuration through environment variables and shared YAML files. This allows consistent plugin setups across development, staging, and production environments without duplicating configuration.
The 118+ plugin count makes OpenClaw one of the most extensible AI agent platforms available. Whether you need to connect to a niche messaging platform, a specific LLM provider, or a custom internal tool, there's likely a plugin — and if not, the API makes building one straightforward.
Getting started with plugins is simple. Most plugins require nothing more than an API key stored in your configuration file. The plugin system handles the rest — automatic capability discovery, runtime loading, and error handling. When the agent identifies a task that matches a plugin's capabilities, it invokes the plugin automatically without needing explicit instructions from the user.
This auto-discovery model is what makes OpenClaw's plugin system uniquely powerful. Unlike traditional plugin architectures where developers must explicitly call plugin functions, OpenClaw's AI agent reasons about which plugin to use based on the task context. Ask it to send a message on Telegram, and it activates the Telegram plugin. Ask it to generate an image, and it routes to the media generation pipeline. No manual wiring required.
📊 See how this tool compares to alternatives
For the official OpenClaw documentation and plugin development guide, visit the OpenClaw docs.
Dig deeper: OpenClaw review · IronClaw review.
Back to all posts
