<<< Back to Index


OpenClaw Capabilities Guide: What You Can Set Up & What It Costs


*Published 2026-03-02 ยท Chrysolambda*




This is a comprehensive inventory of everything OpenClaw can do, what each capability requires, whether it's free, and what API keys or services are involved. Use it to discover capabilities you might be missing.




๐Ÿง  Core LLM (Chat Intelligence)


The brain. Everything else builds on this.


ProviderFree Tier?API Key NeededNotes
Anthropic (Claude)NoANTHROPIC_API_KEYClaude Opus, Sonnet, Haiku
OpenAI (GPT-4, o-series)NoOPENAI_API_KEYGPT-4.1, o3, o4-mini, etc.
Google (Gemini)Yes (generous)GEMINI_API_KEYGemini 2.5 Flash/Pro
OpenRouterPay-as-you-goOPENROUTER_API_KEY300+ models, crypto payments accepted
Ollama (local)โœ… FreeNoneRuns on your GPU; needs VRAM
vLLM (local)โœ… FreeNoneSelf-hosted inference server
Together AIFree trialTOGETHER_API_KEYOpen-source models hosted
MistralFree tierMISTRAL_API_KEYMistral Large, Codestral
Venice AIFree tierVENICE_API_KEYPrivacy-focused, no logging
Amazon BedrockNoAWS credentialsEnterprise; Claude/Llama/etc. via AWS
Hugging FaceFree tierHF_TOKENInference API for open models
NVIDIAFree tierNVIDIA_API_KEYNIM inference endpoints
QwenFreeOAuthAlibaba's models
Moonshot/KimiFree tierMOONSHOT_API_KEYChinese market focus
GLMFree tierAPI keyZhipu AI models
MiniMaxFree tierAPI keyChinese LLM provider
XiaomiFree tierAPI keyMiLM models
QianfanFree tierAPI keyBaidu's model platform
Z.AIVariesAPI keyxAI/Grok models
LiteLLM (proxy)โœ… FreeNone (proxy)Unified gateway to any provider
Vercel AI Gatewayโœ… FreeNone (proxy)Edge proxy for multiple providers
Cloudflare AI GatewayFree tierCloudflare accountCaching/logging proxy layer
Claude Max API Proxyโœ… Free*Claude subscriptionUses your Claude Pro/Max subscription as API

Setup: openclaw onboard or set provider env vars in ~/.openclaw/.env.




๐Ÿ’ฌ Messaging Channels


How OpenClaw talks to you. Each is independent โ€” use one or many.


ChannelFree?RequirementsNotes
Web Chatโœ… FreeNothing extraBuilt-in, localhost
Telegramโœ… FreeBotFather bot tokenCreate bot via @BotFather
WhatsAppโœ… FreeQR code pairingUses your personal WhatsApp; one device
Discordโœ… FreeBot token + serverCreate app at discord.com/developers
Signalโœ… FreeSignal account + linked deviceRequires signal-cli setup
iMessageโœ… FreemacOS + BlueBubbles *or* imsg CLIMac required
Slackโœ… FreeSlack app + bot tokenCreate app at api.slack.com
IRCโœ… FreeServer + nickBuilt-in; TLS + NickServ supported
Matrixโœ… FreeMatrix account + homeserverDecentralized; Element, etc.
Microsoft Teamsโœ… Free*Azure Bot registrationRequires Azure account
Google Chatโœ… Free*Google Workspace + service accountWorkspace admin needed
Mattermostโœ… FreeMattermost server + bot accountSelf-hosted chat
LINEโœ… FreeLINE Messaging API channelLINE Developers console
Nostrโœ… FreeNostr keypairDecentralized protocol
Twitchโœ… FreeTwitch OAuth tokenChat bot integration
Feishu/Larkโœ… FreeFeishu app credentialsByteDance's workspace
Zaloโœ… FreeZalo OA + API credentialsVietnamese messaging
Nextcloud Talkโœ… FreeNextcloud instance + bot accountSelf-hosted
Synology Chatโœ… FreeSynology NAS + Chat packageNAS-based
Tlon (Urbit)โœ… FreeUrbit planet/shipUrbit messaging

Setup: openclaw configure --section channels or edit openclaw.json.




๐Ÿ” Web Search


ProviderFree Tier?API KeyNotes
Brave Search (default)Yes (2,000 queries/mo)BRAVE_API_KEYStructured results
Gemini (Google Search grounding)YesGEMINI_API_KEYAI-synthesized + Google results
Perplexity SonarNo (via OpenRouter)OPENROUTER_API_KEY or PERPLEXITY_API_KEYAI-synthesized answers
Grok (xAI)VariesXAI_API_KEYxAI's search
SearXNG (self-hosted)โœ… FreeNoneVia custom skill; requires Docker

Setup: openclaw configure --section web or set env var.




๐ŸŒ Web Fetch & Scraping


ToolFree?API KeyNotes
web_fetch (built-in)โœ… FreeNoneHTTP GET + Readability extraction
Firecrawl (fallback)Free tier (500 credits)FIRECRAWL_API_KEYAnti-bot scraping, JS rendering
Browser tool (Playwright)โœ… FreeNoneFull browser automation for JS-heavy sites



๐Ÿ–ผ๏ธ Image Generation


MethodFree?RequirementsQuality
OpenRouter free modelsโœ… FreeOPENROUTER_API_KEY + workspace skillFLUX.2 Pro, RiverFlow, etc. โ€” excellent
Gemini (nano-banana-pro)Free tier*GEMINI_API_KEYGemini image gen; may need paid tier
OpenAI DALL-E / gpt-image-1No (~$0.02-0.17/image)OPENAI_API_KEY + bundled skillHigh quality
RecraftFree tierAPI key + ClawHub skillVector + raster; design-focused
Stable Diffusion (local)โœ… FreeGPU + setupVia ComfyUI/A1111 + custom skill

Our setup: OpenRouter free models via workspace skill (skills/openrouter-imagegen/). Works great, zero cost.




๐Ÿ‘๏ธ Image Recognition / Vision


MethodFree?RequirementsNotes
Built-in image toolDepends on modelagents.defaults.imageModel configUses your LLM provider's vision model
Claude VisionNoANTHROPIC_API_KEYExcellent; auto-paired with Claude models
GPT-4 VisionNoOPENAI_API_KEYStrong general vision
Gemini VisionYes (free tier)GEMINI_API_KEYGood; generous free quota
Ollama vision modelsโœ… FreeLocal GPULLaVA, Moondream, etc.

Setup: Set agents.defaults.imageModel in config, or it auto-pairs from your primary model's provider.




๐ŸŽค Speech-to-Text (Transcription)


Transcribes voice notes sent to you via messaging channels.


ProviderFree?RequirementsNotes
OpenAI Whisper APINo (~$0.006/min)OPENAI_API_KEYCloud; high accuracy; bundled skill
DeepgramFree tier (limited)DEEPGRAM_API_KEYCloud; fast; built-in provider
Whisper CLI (local)โœ… Freewhisper binary + GPURuns locally; bundled skill
Sherpa-ONNX (local)โœ… Freesherpa-onnx binaryLightweight offline STT

Setup: For cloud: set API key + enable under tools.media.audio. For local: install the binary.




๐Ÿ”Š Text-to-Speech (TTS)


ProviderFree?RequirementsNotes
Edge TTS (default)โœ… FreeNothingMicrosoft's neural voices; no API key
OpenAI TTSNo (~$0.015/1K chars)OPENAI_API_KEYHigh quality; gpt-4o-mini-tts
ElevenLabsFree tier (10K chars/mo)ELEVENLABS_API_KEYBest quality; voice cloning; expressive
Sherpa-ONNX TTS (local)โœ… Freesherpa-onnx binaryFully offline; bundled skill
sag (ElevenLabs CLI)Free tierELEVENLABS_API_KEYMac-style say UX; bundled skill

Setup: Enable via messages.tts.auto in config. Edge TTS works out of the box.




๐Ÿ“ž Voice Calls


ProviderFree?RequirementsNotes
TwilioNo (pay-per-minute)Twilio account + phone numberOutbound + inbound calls
TelnyxNo (pay-per-minute)Telnyx account + numberCall Control v2
PlivoNo (pay-per-minute)Plivo account + numberVoice API

Setup: Install voice-call plugin, configure under plugins.entries.voice-call.config.




๐ŸŽต Music Generation


MethodFree?RequirementsNotes
Diffrythm / Tencent (via ClawHub)โœ… FreeClawHub skill + inference.sh CLIText-to-music; instrumentals or vocals
ACE-Step 1.5โœ… FreeClawHub skillMusic generation API

Setup: clawhub install ai-music-generation




๐ŸŽฌ Video Capabilities


CapabilityFree?RequirementsNotes
Camera capture (photo/video)โœ… FreePaired iOS/Android/macOS nodeShort clips via nodes โ†’ camera_clip
Screen recordingโœ… FreePaired nodenodes โ†’ screen_record
Video frame extractionโœ… Freeffmpeg installedBundled skill; extract stills from video
Video generationVia external APIVariesNo built-in; use skills for Runway, Kling, etc.



๐Ÿ–ฅ๏ธ Browser Automation


FeatureFree?Requirements
Playwright browserโœ… Freebrowser.enabled: true (default)
Chrome Extension relayโœ… FreeChrome + OpenClaw extension
Multi-profile supportโœ… FreeConfig; up to ~100 profiles
Screenshotsโœ… FreeBrowser enabled
PDF generationโœ… FreeBrowser enabled

Setup: Works out of the box. Install Playwright browsers with npx playwright install.




๐Ÿ“ฑ Node Network (Device Control)


Pair phones, Macs, or other machines for remote control.


CapabilityFree?Requirements
Camera snap/clipโœ… FreePaired node with camera permission
Screen recordingโœ… FreePaired node with screen permission
Notificationsโœ… FreePaired macOS node
Remote command executionโœ… FreePaired node with exec permission
Locationโœ… FreePaired node with location permission
Canvas (A2UI)โœ… FreePaired node

Setup: Pair via openclaw nodes pair or the macOS/iOS app.




๐Ÿ“ง Email


MethodFree?RequirementsNotes
gog (Google Workspace CLI)โœ… FreeGoogle account + OAuthGmail, Calendar, Drive, Sheets, Docs
himalayaโœ… FreeIMAP/SMTP credentialsAny email provider; CLI-based
AgentMailFree tierAPI keyDedicated agent email address



๐Ÿ“… Calendar & Tasks


MethodFree?RequirementsNotes
gog (Google Calendar)โœ… FreeGoogle OAuthFull calendar management
Apple Reminders (remindctl)โœ… FreemacOSBundled skill
Things 3โœ… FreemacOS + Things appBundled skill
Trelloโœ… FreeTrello API keyBoards, lists, cards
Linearโœ… FreeLinear API keyClawHub skill
Notionโœ… FreeNotion API keyPages, databases, blocks



๐Ÿ“ Notes & Knowledge


MethodFree?Requirements
Obsidian (local vaults)โœ… FreeObsidian + obsidian-cli
Apple Notes (memo CLI)โœ… FreemacOS
Bear Notes (grizzly CLI)โœ… FreemacOS + Bear app
Notionโœ… FreeNotion API key



๐Ÿ  Smart Home


MethodFree?RequirementsNotes
Home Assistantโœ… FreeHA instance + REST APIClawHub skill; control devices, scenes
Philips Hue (OpenHue)โœ… FreeHue Bridge + openhue CLIBundled skill; lights & scenes
Eight Sleepโœ… FreeEight Sleep pod + eightctlBundled skill; sleep temperature
Sonos (sonoscli)โœ… FreeSonos speakers on LANBundled skill; playback control



๐ŸŽง Media & Entertainment


MethodFree?Requirements
Spotify (spotify_player/spogo)โœ… Free*Spotify account + CLIPlayback/search
BluOS (blucli)โœ… FreeBluOS speakersDiscovery, playback, volume
GIF search (gifgrep)โœ… Freegifgrep CLISearch + download GIFs
Audio visualization (songsee)โœ… Freesongsee CLISpectrograms from audio



๐Ÿฆ Social Media


PlatformFree?Requirements
X/Twitter (xurl)โœ… Free*X API credentials (OAuth)Post, reply, search, DMs
Blog monitoring (blogwatcher)โœ… Freeblogwatcher CLIRSS/Atom feed watching



๐Ÿ” Secrets Management


MethodFree?Requirements
1Password (op CLI)No (subscription)1Password + op CLI
Environment variablesโœ… Free~/.openclaw/.env file
Config secretsโœ… Freeopenclaw.json SecretRef



๐Ÿ“„ Document Processing


MethodFree?Requirements
PDF editing (nano-pdf)โœ… Freenano-pdf CLINatural-language PDF edits
Summarize (URLs/podcasts/video)โœ… Freesummarize CLITranscripts + summaries
Web content extractionโœ… FreeBuilt-in web_fetchHTML โ†’ markdown



๐Ÿค– Multi-Agent Orchestration


FeatureFree?Requirements
Sub-agents (sessions_spawn)โœ… Free*Same LLM provider costsSpawn isolated agents for tasks
ACP agents (Claude Code, Codex)VariesRespective API keysDelegate coding to specialized agents
Cron jobsโœ… FreeGateway runningScheduled agent tasks
Heartbeatsโœ… FreeGateway runningPeriodic proactive check-ins
Cross-session messagingโœ… FreeMultiple agents configuredAgent-to-agent communication



๐Ÿช Hooks & Webhooks


Event-driven automation that runs when things happen โ€” no polling required.


Internal Hooks


Hooks are TypeScript handlers that fire on agent/gateway events. All free, no API keys.


HookEventWhat It Does
๐Ÿ’พ session-memory (bundled)command:newSaves session context to memory files before reset
๐Ÿ“Ž bootstrap-extra-files (bundled)agent:bootstrapInjects extra workspace files (e.g. monorepo AGENTS.md)
๐Ÿ“ command-logger (bundled)commandAudit log of all commands to JSONL file
๐Ÿš€ boot-md (bundled)gateway:startupRuns BOOT.md instructions when gateway starts

Available events to hook into:


EventWhen It Fires
command:new / command:reset / command:stopSlash commands issued
message:receivedInbound message from any channel
message:sentOutbound message delivered
agent:bootstrapBefore workspace files are injected
gateway:startupAfter gateway starts and channels connect
tool_result_persistBefore tool results are saved (sync, plugin API)

Popular community hook patterns:



Setup:


openclaw hooks list              # see all hooks
openclaw hooks enable <name>     # turn one on
openclaw hooks info <name>       # check requirements

Custom hooks go in ~/.openclaw/hooks/my-hook/ or <workspace>/hooks/my-hook/ with a HOOK.md + handler.ts. Can also be distributed as hook packs (npm packages) via openclaw hooks install <spec>.


Webhooks (External HTTP Triggers)


Let external services trigger OpenClaw actions via HTTP POST. Free, requires shared secret token.


EndpointWhat It Does
POST /hooks/wakeEnqueue a system event + trigger immediate heartbeat
POST /hooks/agentRun an isolated agent session with a custom prompt

Popular webhook patterns:



Config:


{
  hooks: {
    enabled: true,
    token: "your-shared-secret",   // required
    path: "/hooks",                 // default
  }
}

All webhook requests require Authorization: Bearer <token> header.




โฐ Automation & Scheduling


FeatureFree?Requirements
Cron (built-in)โœ… FreeGateway runningFull cron syntax
Heartbeat pollingโœ… FreeGateway runningConfigurable interval
Broadcast groupsโœ… FreeMultiple channelsSend to many channels at once



๐Ÿ”ง System & Development


FeatureFree?Requirements
Shell execution (exec)โœ… FreeBuilt-inRun any command
File I/O (read/write/edit)โœ… FreeBuilt-inFull filesystem access
GitHub (gh CLI)โœ… Freegh auth + PATIssues, PRs, CI, code review
GitLabโœ… FreePATClawHub skill
MCP servers (mcporter)โœ… Freemcporter CLIConnect any MCP tool server
tmux controlโœ… Freetmux installedRemote-control terminal sessions
Google Places (goplaces)NoGoogle API keyPlace search, details, reviews



๐ŸŒค๏ธ Miscellaneous


CapabilityFree?Requirements
Weatherโœ… FreeNothingwttr.in / Open-Meteo; bundled skill
RTSP/ONVIF cameras (camsnap)โœ… FreeCamera URLIP camera capture
Food delivery tracking (ordercli)โœ… FreeFoodora accountOrder status
macOS UI automation (peekaboo)โœ… FreemacOS + peekaboo CLIScreenshot + automate macOS apps
Location sharingโœ… FreeTelegram/WhatsApp/MatrixParse inbound locations
ClawHub (skill store)โœ… Freeclawhub CLI5,700+ community skills



๐Ÿšซ What OpenClaw Does NOT Do (Natively)


These require external services, custom skills, or workarounds:


CapabilityWorkaround
Video generation (Sora, Runway, Kling)Custom skill + API key for the service
3D model generationCustom skill + external API
Real-time voice conversation (like GPT-4o voice)Voice call plugin (Twilio/Telnyx) gets close
OCR (document scanning)Vision model or Tesseract via exec
Translation serviceLLM handles it natively; DeepL skill on ClawHub
Database queriesVia exec (psql, sqlite3, etc.) or MCP server
Payment processingNot built-in; would need custom integration
Continuous audio monitoringNot supported; snapshot-based only



๐Ÿ’ฐ Cost Summary: The "Totally Free" Stack


You can run a fully functional OpenClaw with zero ongoing costs using:


LayerFree Option
LLMOllama (local) or Gemini free tier
MessagingTelegram + Web Chat
SearchBrave (2K queries/mo free) or Gemini
Image GenOpenRouter free models
VisionGemini free tier or Ollama (LLaVA)
TTSEdge TTS (no API key)
STTWhisper CLI (local)
BrowserPlaywright (built-in)
All toolsBuilt-in (exec, file I/O, memory, cron)

Total: $0/month (assuming you have hardware for Ollama)




๐Ÿ’Ž The "Best Experience" Stack


For maximum capability, expect ~$20-50/month:


LayerRecommendedCost
LLMClaude Opus + Gemini Flash fallback~$20-40/mo
SearchBrave free tier + Gemini groundingFree
Image GenOpenRouter free modelsFree
VisionClaude Vision (auto-paired)Included
TTSElevenLabs (free tier) + Edge fallbackFree-$5/mo
STTDeepgram free tierFree
Voice CallsTwilio~$1-5/mo



*Sources: OpenClaw docs (local), clawhub.com, ClawHub ecosystem research, bundled skill inventory.*



<<< Back to Index