Agenten-Kompass

AGENT PROFILES

OpenClaw

OpenClaw Foundation · General-purpose agents

Self-hosted, model-agnostic assistant with a persistent gateway to chat channels and tools.

Research Organize office work Automate

OpenClaw · Self-hosted agent runtime and gateway

Product and access

As a self-hosted agent runtime, OpenClaw connects models, chat channels and tools through a persistent gateway.

Provider

The independent US 501(c)(3) foundation stewards OpenClaw, employs the core team and signs releases. Donations fund development; the software remains MIT licensed.

Operator
OpenClaw Foundation
Organization
Nonprofit foundation
Location
USA
GitHub stars
391,505 · Reviewed
License
MIT
Provider sourcesSource 1 Source 2 Source 3 Reviewed
01

Core features

9 features from provider sources · checked on 9 October 2026. Availability and limits are listed below.

  • Self-hosting

    Runs on your own infrastructure or a shared team gateway.

    Source 1

  • Chat channels

    Connects services including WhatsApp, Telegram, Signal, Slack, and WebChat.

    Source 1

  • Sessions and memory

    Runs cross-channel sessions with persistent context.

    Source 2

  • Tools

    Agents can use configured actions such as a browser or command execution.

    Source 3

  • Skills

    Markdown instructions provide reusable workflows.

    Source 4

  • Extensions

    Plugins add channels, models, tools, skills, and runtime features.

    Source 3

  • Automation

    Scheduled workflows and background work can continue later.

    Source 3

  • Hooks

    Lifecycle and tool calls can be observed or controlled through extensions.

    Source 5

  • Shared team gateway

    Teams can open and steer sessions together; live presence shows participation.

    Source 7
02

Requirements and limits

  • Requirement

    Your own runtime

    OpenClaw must be set up by the user; model providers and desired channels are configured separately.

    Source 6

  • Limitation

    Tool access

    Actions depend on the active tool profile, access policies, channel permissions, and available plugins.

    Source 3

  • Requirement

    Set up each channel

    Each chat channel requires its own connection and access configuration; for WhatsApp, the guide recommends approving senders.

    Source 6

Operation & oversight

Operational details for this profile have not yet been researched.

03

Practical tasks

Editorial examples based on documented features. No product tests performed.

01

Research through chat

Your input
Name a topic and desired sources in the connected channel.
Possible result
A configured agent can use web tools and reusable research Skills.
What to check
Check channel access, tool policy, sources, and outputs at your own gateway.
02

Recurring reminder

Your input
Set up a schedule and a limited briefing task.
Possible result
OpenClaw can report scheduled background work back to the main session.
What to check
Check the schedule, recipients, permissions, and runtime behavior in your own installation.
EDITORIAL EXAMPLE

Contact a custom assistant for approved workflows through a chat channel.

Your possible prompt

Compare three providers for [task]. Use their official documentation. Show differences, sources, and open questions in a table.

Getting-started idea based on documented capabilities. Not an executed product test. Adapt the task, data, and permissions.

04

Setup and pricing

Setup

Technical setup and model access

Prices and included usage

Pricing checked

MIT software is provided without a subscription or paid hosted tier. Model access and your own infrastructure are funded separately.

OpenClaw
USD 0 software license

Self-hosted runtime with configurable models, channels and tools.

Price source
  • Self-hosting adds infrastructure and, where applicable, model costs. OpenAI's price list is one example of separately billed model access: https://developers.openai.com/api/docs/pricing

Pricing details

05

Sources and evidence

Official product and feature documentation reviewed again on 9 October 2026. Examples are editorial use suggestions; no hands-on product tests were performed here.

Editorial team: Michael Meierhoff · sources reviewed on . Provider information is not independent evidence of performance.

Read profile as Markdown

How we assess the sources