How OpenClaw Went from Launch to 1,299 Repos in 8 Weeks

OpenClaw launched in late January 2026. Eight weeks later, PT-Edge tracks 1,299 repositories across 13 categories built around it. That's not a project growing — it's an ecosystem forming in real time.

We pulled the GitHub creation dates for the top 25 projects and mapped the timeline. A clear pattern emerged: not random growth, but a structured sequence of layers being built in a specific order. The same pattern is likely playing out in other AI ecosystems, just slower.

The pre-existing layer: projects that pivoted to OpenClaw

Three of the most popular projects in the OpenClaw ecosystem weren't built for OpenClaw at all. They existed before it launched and adapted to integrate it.

Super Agent Party started as an all-in-one AI companion in March 2025 — nearly a year before OpenClaw. AionUi was a desktop cowork app for coding agents since August 2025. n8n-skills was an n8n workflow toolkit from October 2025. All three added OpenClaw support and became major ecosystem players after the launch.

This matters because it's a signal of ecosystem pull. When existing projects independently decide to integrate a new platform, it means the platform is generating real demand — not just excitement. AionUi now has 18,636 stars and 1618 commits in the last 30 days, making it the most active project in the entire OpenClaw ecosystem.

Week 1-2: Localisation and deployment (late January)

Within days of OpenClaw going viral, three things happened simultaneously: someone translated it to Chinese, someone built a deployment wrapper, and someone gave it a voice.

This first wave reveals what people need immediately when a new AI tool goes viral: make it work where I already am. Chinese localisation (3,138 stars, syncs hourly with the main repo), one-click deployment (openclaw-coolify, 80% fork rate — people fork it as a deployment template), and voice activation (openclaw-assistant for Android).

Week 3-6: Platform buildout (February)

February was when the ecosystem went from "people trying OpenClaw" to "people building on OpenClaw." Fifteen of the top 25 projects were created in this window.

Clients for every surface:

Android (openclaw-android), desktop with multi-platform messaging (Nexu), visual dashboard (ClawPanel), Feishu integration (openclaw-feishu). The pattern: meet people where they already work. Especially notable is the Chinese workplace integration — Feishu and WeChat connectors appeared fast because the Chinese community was already building.

Infrastructure and skills:

ClawRouter (5,432 stars, created February 3) is an LLM router supporting 41+ models with sub-millisecond routing — the kind of middleware that turns a single-model agent into a multi-model platform. AlphaClaw (11,234 downloads/month) solved the deployment pain point and is the only infrastructure project with meaningful package adoption.

The first domain-specific agents also appeared in this wave. ClawBio — bioinformatics-native AI agent skills. HyperLiquid-Claw — crypto trading via OpenClaw. These are early but they signal that the platform is expanding beyond general-purpose chat.

Week 6-8: Ecosystem maturation (March)

March brought the layer that makes an ecosystem sustainable: education, documentation, and orchestration.

hello-claw from Datawhale China is the first systematic tutorial — "adopt from scratch and build your first claw." openclaw-master-skills curates 339+ skills, updated weekly. There's even a full book about OpenClaw, written using OpenClaw itself.

Multi-agent orchestration appeared: ClawTeam-OpenClaw for swarm coordination, OpenCrew for Slack-based agent teams. And MetaClaw (805 stars, 171 commits in 30 days) introduced a self-evolving agent that learns from conversation.

The Chinese community built the adoption layer

One of the most striking patterns in the timeline is that Chinese-language projects didn't just participate — they built the onboarding infrastructure that the core project didn't provide.

OpenClawChineseTranslation (3,138 stars) appeared on January 30, three days after launch. awesome-openclaw-usecases-zh (2,074 stars) documents 40 real-world use cases adapted for the Chinese ecosystem. hello-claw, openclaw-docs, openclaw-feishu, CountBot, wechatbot — tutorials, documentation, platform integrations, all in Chinese.

This isn't translation. It's original work: use case guides, platform-specific integrations (Feishu, WeChat, 微博), and educational content tailored for a different developer culture. When an ecosystem independently grows communities in multiple languages building original tooling, it's a strong signal that something durable is being built.

What the adoption data actually shows

Stars tell one story. Downloads and fork rates tell another.

The ecosystem has 67,379 stars across the featured projects, but only 15 repos have any package downloads at all. Most of these projects aren't libraries you pip install — they're applications you clone and run, or configs you fork and customise. That's a different adoption model than something like PyTorch or LangChain.

The fork rates confirm this. n8n-skills has a 18% fork rate — people fork it to customise their own skill sets. openclaw-master-skills has a 22% fork rate. The deployment template openclaw-coolify has an 80% fork rate — almost everyone who finds it forks it to run their own instance.

The one project with clear package adoption is AlphaClaw: 11,234 npm downloads per month. It solved the "how do I actually keep this running" problem and people install it.

The pattern

If you step back from the specifics, the OpenClaw ecosystem followed a sequence:

1. Existing projects pivot in — established tools add support for the new platform, bringing their users with them
2. Localisation and deployment — make it work where people already are (language, platform, infrastructure)
3. Platform buildout — clients, skills, middleware, domain-specific agents
4. Education and orchestration — tutorials, documentation, multi-agent coordination, the pieces that make it accessible to non-early-adopters

This happened in 8 weeks. Not every AI ecosystem will move this fast — MCP took longer, the agents space is still diffuse — but the sequence of layers is likely similar. The first things built are always about access (run it here, read it in my language). The last things built are always about mastery (orchestrate multiple agents, write a book).

We'll be watching for this pattern in other ecosystems as they form.

PT-Edge tracks 226,000+ AI repositories across 18 domains, including 1,299 in the OpenClaw ecosystem. Every project linked above has a full detail page with daily-updated metrics. Browse the full ecosystem: client applications, skill marketplace, agent templates, infrastructure.