Getting Started
What is goClaw
Overview of the goClaw AI agent platform — what it is, how it works, and why it's different.
goClaw is Indigo's deployable AI agent platform. It packages the NanoClaw agent runtime, a built-in CRM, multi-channel communications, self-evolving knowledge, and a permission system into a single managed product.
Tagline: Deploy an agent. It learns the rest.
The platform is built on ClawRM — Indigo's internal agent engine — and is designed to be lightweight by default. No vector databases, no Kubernetes, no infra team. SQLite and Markdown all the way down.
What Makes goClaw Different
Lightweight by design
Most agent platforms require substantial infrastructure: vector databases, orchestration frameworks, workflow builders, and dedicated DevOps. goClaw runs on SQLite, Markdown, and a single Node.js process. Docker is optional. EC2 t3.small is enough for most tenants.
Six interlocking pillars
goClaw isn't a single-purpose tool. It's six components that work together:
| Pillar | What it does |
|---|---|
| NanoClaw Core | Agent runtime — scheduling, execution, tool use |
| Built-in CRM | Contacts, threads, notes — unified across channels |
| Multi-Channel Comms | Email, SMS, Telegram, WhatsApp |
| Self-Evolving Knowledge | Markdown + FTS5 + nightly research pipeline |
| Permission System | Group isolation, sandbox execution, container mode |
| Your Tools & Data | MCP tool integration, custom datasources |
The agent teaches itself
This is the core differentiator. When goClaw encounters a gap — a question it can't answer, a procedure it doesn't know — it files a curiosity item. A nightly research pipeline processes the queue: web search → LLM synthesis → categorization → deduplication → new knowledge files.
After 30 days, an active agent typically has 60+ knowledge files it wrote itself. It gets better every day without you touching it.
vs. OpenClaw
OpenClaw is a heavy, general-purpose agent runtime (247K GitHub stars). Powerful but complex. goClaw is lightweight by design — 5-minute CLI setup vs. weeks of configuration.
vs. Clay / Apollo / Outreach
These tools require manual workflow building. goClaw is autonomous: plug in your tools, it learns its purpose, it executes. The agent makes decisions, not just follows steps.
vs. Custom Builds
Building your own agent means wiring up CRM, channels, knowledge, permissions, and learning from scratch. goClaw gives you all six pillars out of the box.
Name & Identity
Name origin: "indiGO" + "Claw" — the GO in Indigo, plus the claw agent motif. A lobster claw grips with tenacity, pursues autonomously, and doesn't let go once locked on. The visual asymmetry — lowercase "go" + uppercase "Claw" — mirrors the Indigo brand style.
Domain: goclaw.getindigo.ai (goclaw.ai was taken)
Who is it for?
goClaw is built for teams that want:
- Autonomous outreach or support without manual workflow configuration
- An agent that improves with use, not one that needs constant reprogramming
- A full stack (CRM + channels + knowledge + permissions) without stitching together five tools
- Lightweight infrastructure they can actually understand and debug
If you're a developer or technical founder running GTM, support, or sales ops — goClaw is built for you.