<2-3 sentences summarising what the project addresses and solves, the target audience, and the benefit.> klaw.sh provides enterprise AI agent orchestration, akin to kubectl for AI workloads. It simplifies the management, monitoring, and scaling of AI agents, offering a single binary deployment for rapid setup and operation, benefiting teams running AI agents in production.

How It Works

klaw.sh employs a Kubernetes-inspired architecture, managing AI agents within isolated namespaces. It utilizes a single Go binary for deployment and operation, featuring a controller-worker model for distributed task execution. This approach offers simplified deployment, enhanced visibility, and robust scaling for AI agent fleets.

Quick Start & Requirements

• Primary install: curl -fsSL https://klaw.sh/install.sh | sh
• Prerequisites: API keys for LLM providers (e.g., Anthropic, OpenRouter, each::labs), and Slack tokens (SLACK_BOT_TOKEN, SLACK_APP_TOKEN) for Slack integration.
• Run: klaw start to launch the full platform (Slack + scheduler + agents). klaw chat for interactive terminal chat.
• Links: Documentation, Quick Start, Slack Control, Architecture.

Highlighted Details

• kubectl-style CLI for agent management (list, describe, logs, create, delete).
• Slack integration enables agent control, status checks, and quick queries directly from Slack.
• Namespace isolation provides logical separation for agents, including scoped secrets and tool permissions.
• Supports 300+ models via routers like each::labs, or direct integrations with OpenAI, Anthropic, Google, Azure, Ollama, and OpenAI-compatible endpoints.
• Built-in cron scheduler for automated agent runs and distributed task dispatching to worker nodes.

Maintenance & Community

• Developed by each::labs.
• Community support available via Discord.
• Bug reports and feature requests can be submitted via GitHub Issues.
• Updates and announcements on Twitter.

Licensing & Compatibility

• License: Source-available under the each::labs License.
• Compatibility: Free for internal business use, personal projects, and consulting. A license is required for multi-tenant SaaS or white-label distribution. See LICENSE for full details. Enterprise licensing inquiries can be directed to enterprise@eachlabs.ai.

Limitations & Caveats

Non-containerized agents lack filesystem sandboxing and execute under the host user's permissions. For true process and filesystem isolation, running agents within containers (e.g., using Podman) is recommended.