Skip to content
JuicyAgents

Supported AI · OpenClaw Foundation

What is OpenClaw?

OpenClaw is a free, open-source AI assistant that runs on your own computer and talks to you in Telegram, Slack, WhatsApp or its web dashboard. You connect your own AI model login or API key. It can browse the web, run scheduled automations and use skills, which are folders with a SKILL.md file.

Last checked · Sources

Who can use it

  • Free and open source (MIT), with no hosted tier
  • Runs on macOS, Windows, Linux or WSL2
  • Claude Code or Codex login, or a model API key
  • Node.js 24.16.0 or newer on 24.x, or 26.1.0 or newer; 24.22.1 or 26.12.1 and newer advised for security fixes
  • Some setup in a terminal

OpenClaw at a glance

Each fact links to the page it comes from, with the day we last checked that page.

OpenClaw facts, with their sources and the day we checked them
FactWhat the docs say
Where it runsYour own computer or server, with a terminal UI, a web dashboard and chat apps such as Telegram and SlackSource: OpenClaw getting started(opens in a new tab)Checked
Connect appsGmail, Google Calendar, Sheets, Docs and Drive search through the bundled gog skill; a built-in browser and web search; Slack, WhatsApp and other chat channelsSource: OpenClaw gog skill(opens in a new tab)Checked
SchedulesBuilt-in automations with cron times, intervals or one-time runs; the Gateway must be running for them to fireSource: How OpenClaw automations work(opens in a new tab)Checked
PlanFree and open source (MIT), with no hosted tier; you pay your model provider or reuse a Claude Code or Codex loginSource: OpenClaw repository(opens in a new tab)Checked
AvailabilityNo country limits in the official docs; your model provider's terms applySource: OpenClaw repository(opens in a new tab)Checked

What OpenClaw can do

  • Skills from SKILL.md folders

    OpenClaw follows the AgentSkills format: each skill is a folder with a SKILL.md file. Save one in ~/.openclaw/workspace/skills/ and call it with /skill and its name. Read third-party skills first, because the docs call them untrusted code.

    Example: /skill meeting-recap Turn these notes into decisions and actions.

  • Automations on a schedule

    Automations are OpenClaw's built-in scheduler. Create a recurring job with openclaw automations create and a cron time, and send the result to a chat channel. Schedules fire only while the Gateway is running.

    Example: Every weekday at 7am, summarize overnight updates and post them to Slack.

  • Browser, web search and Google apps

    OpenClaw can control its own separate browser profile and search the web to read public pages. The bundled gog skill connects Gmail, Calendar, Sheets, Docs and Drive search through your own Google Cloud OAuth client.

    Example: Search the web for this company's latest news and list the sources with dates.

Get started with OpenClaw

  1. Install OpenClaw with the official script: curl -fsSL https://openclaw.ai/install.sh | bash on macOS, Linux or WSL2, or the PowerShell command from the install guide on Windows.
  2. Let onboarding run Quick start. It reuses a Claude Code or Codex login or an API key, checks it with a real reply and opens the web dashboard.
  3. Press Ctrl+C, run openclaw gateway install so it keeps running in the background, then check openclaw gateway status. Before you connect real accounts, set tools.exec.mode to ask so commands need your approval.
  4. Pick a skill or recipe below. Save a skill as ~/.openclaw/workspace/skills/<skill-name>/SKILL.md, start a new session and call it with /skill and the skill's name.
  5. For scheduled work, keep the Gateway running, create the job with openclaw automations create, and check openclaw automations runs before you rely on it.

What to check before you trust it

  • By default OpenClaw runs commands without asking. Set tools.exec.mode to ask or allowlist before it can reach real accounts or files.
  • Agents with message access can send across conversations and channels by default, so keep prompts draft-only and review anything before it is sent.
  • Treat third-party skills as untrusted code and read each SKILL.md before you enable it.
  • Web pages, emails and documents it reads can carry hidden instructions. The docs advise against older or smaller model tiers for agents that use tools or read inboxes.
  • Run openclaw security audit to catch risky settings, and test with a fictional sample before you use real data.

We have not run these setups ourselves. Recipes are written from public OpenClaw documentation, and their examples are illustrative.

Recipes built with OpenClaw (7)

A recipe is a step-by-step setup you build in OpenClaw yourself. Each one shows what you provide, what you get, and a setup prompt to copy. Examples are illustrative, not recorded results.

Skills you can add to OpenClaw (70)

A skill is a folder of instructions that OpenClaw can use for a specific task. Add it once, then ask for the task in your own words.

  • Sales call follow-up

    Turn call notes into a recap email that promises only what was agreed, proposed CRM changes and a task list with owners and dates.

    • Codex
    • Claude Code
    • Claude app
    • OpenClaw
    • Hermes Agent
  • Account reconciliation

    Compare accounting records with a clear audit trail, separate uncertain matches, and explain remaining differences.

    • Codex
    • Claude Code
    • OpenClaw
    • Hermes Agent
  • Research synthesis

    Combine supplied sources into a clear answer that preserves evidence, disagreement, and uncertainty.

    • Codex
    • Claude Code
    • OpenClaw
    • Hermes Agent
  • Document field extraction

    Extract requested fields into a source-linked table without guessing missing or unreadable values.

    • Codex
    • Claude Code
    • OpenClaw
    • Hermes Agent
  • Prospect fit research

    Check a named company against your customer criteria and return a sourced fit brief.

    • Codex
    • Claude Code
    • OpenClaw
    • Hermes Agent
  • Business process documentation

    Turn team notes into a repeatable process with clear owners, decisions, exceptions, and completion checks.

    • Codex
    • Claude Code
    • OpenClaw
    • Hermes Agent
  • AI job search

    Find current job openings, drop duplicates and expired posts, and shortlist them against your must-haves and real experience.

    • Codex
    • Claude Code
    • Claude app
    • OpenClaw
    • Hermes Agent
  • Applicant tracker from email

    Log job applications from Gmail or Outlook in a Google Sheet or CSV, and see which applicants are waiting for your reply.

    • Codex
    • Claude Code
    • OpenClaw
    • Hermes Agent
  • Article-to-Visuals

    Map an approved article to a cover, three illustrations and an optional six-card carousel, exported as editable SVG, PNG and PDF files.

    • Codex
    • Claude Code
    • OpenClaw
    • Hermes Agent
  • Audio edit plan

    Create a cut and keep plan from a transcript while protecting the speaker’s meaning.

    • Codex
    • Claude Code
    • OpenClaw
    • Hermes Agent
  • Budget versus actual review

    Compare budget and actual figures on the same basis and explain only supported differences.

    • Codex
    • Claude Code
    • OpenClaw
    • Hermes Agent
  • Bug diagnosis

    Trace a failure to its first broken invariant, test competing causes and verify a minimal authorized fix.

    • Codex
    • Claude Code
    • OpenClaw
    • Hermes Agent
Browse all 70 skills

Questions about OpenClaw

Is OpenClaw free?
Yes. OpenClaw is open source under the MIT license and has no paid tier or hosted service. You pay for the AI model it uses, through an API key or an existing Claude Code or Codex login.
Can OpenClaw use JuicyAgents skills?
Yes. OpenClaw follows the AgentSkills format, and our skills have the name and description it needs. Save the SKILL.md in ~/.openclaw/workspace/skills/<skill-name>/, start a new session and call it with /skill and the skill's name.
Can OpenClaw read my Gmail?
Through the bundled gog skill. It needs the gog command-line tool and your own Google Cloud OAuth client, then you add your account with the services you allow, such as Gmail and Calendar. The same skill can send email, so keep recipes draft-only.
Does OpenClaw ask before it acts?
Not by default. Commands run without approval unless you set tools.exec.mode to ask or allowlist, and agents with message access can send messages. Change these settings before you connect real accounts.
Can OpenClaw run tasks on a schedule?
Yes, with automations. Run openclaw automations create with a cron time, such as "0 7 * * *" for 7am every day, and choose where the result goes. Schedules fire only while the Gateway is running.

Get new templates and agents by email

New free templates, and AI agents you can set up in the AI you already pay for.

Email us to subscribe

Sign up for occasional JuicyAgents emails about new agent guides and useful workflows. We will use your email only for these updates. You can unsubscribe from any email. Privacy policy

Sources

The pages behind OpenClaw’s facts and setup steps, with the day we last checked each one.

All sources (26)

Compare OpenClaw with

Compare all 13 AIs side by side