EN
Back to the archive

The archive · AI & Models · Product decision · 2025–2026

HappyCapy bets the browser becomes an agent-native computer; PH #1, $1M ARR in a month

Trickle's team built an OpenClaw-style agent workspace that runs in a browser tab — #1 on Product Hunt, #1 for its week, and $1M ARR inside its first month.

HappyCapy (Trickle Pte. Ltd.)

The betThat people would adopt an "agent-native computer" — a GUI cloud sandbox with Claude Code in the browser — because self-hosting OpenClaw agents was too engineering-heavy.Scaling

What the business is

HappyCapy sells a browser-based AI workspace: an isolated cloud sandbox where agents powered by Claude Code and 150+ other models code, browse, run files and complete multi-step tasks, with a skills marketplace for non-technical users.

Starting capitalOver $10M in seed and pre-Series A funding from HongShan Seed Fund, Planetree VC, Monolith and Baidu VC (hokai.io review, 2026-08-10).

How it started

HappyCapy came out of the team behind Trickle, an AI vibe-coding canvas, led by Jarod Xu, an early Claude Code power user who pushed vibe coding inside his own team from 2025. When OpenClaw went viral in early 2026, the team saw both its capability and its adoption barrier — self-hosting — and built HappyCapy as an "OpenClaw alternative, in your browser," with beta access from January 27, 2026.

What happened

HappyCapy launched on Product Hunt on 2026-02-11 and finished #1 Product of the Day with 1.4k upvotes and 590 comments, then #1 Product of the Week (Feb 9–15) and #2 on February's monthly board with 1,244 votes. It reported $1M ARR within its first month and raised over $10M in seed and pre-Series A funding from HongShan Seed Fund, Planetree VC, Monolith and Baidu VC; by August 2026 the product had expanded to an iOS app, 30k+ skills and 150+ built-in models, with founder Jarod Xu publicly describing the move from Trickle to HappyCapy as a deliberate bet that the computer itself becomes agent-native.

How it ended up

Still scaling as of September 2026: HappyCapy runs free, Pro ($20/month) and Max tiers, ships weekly updates to its skills marketplace, and sells itself as "the agent-native computer for the rest of us" — competing with established agentic platforms like Manus and Devin for the same jobs.

Background

HappyCapy is a browser-based "agent-native computer" built by the team behind Trickle, an AI vibe-coding canvas. Founder Jarod Xu was an early Claude Code power user who pushed vibe coding inside his team from 2025; when OpenClaw went viral in early 2026, he saw both the capability and the adoption barrier — users had to install, configure and keep a machine running. HappyCapy's answer was a cloud sandbox with a GUI: open a tab, and Claude Code has a live desktop it can operate, with 150+ models and a skills marketplace behind it.

The launch was unusually strong. HappyCapy hit #1 Product of the Day on Product Hunt on 2026-02-11 with 1.4k upvotes and 590 comments, took #1 for its week (Feb 9–15, 2026) and finished #2 on February's monthly board with 1,244 votes. The company reported $1M ARR within its first month and raised over $10M in seed and pre-Series A funding from HongShan Seed Fund, Planetree VC, Monolith and Baidu VC.

By August 2026 the product had expanded to an iOS app with 30k+ skills and 150+ built-in models, with free, Pro ($20/month) and Max tiers. The move from Trickle to HappyCapy was an explicit repositioning: instead of another AI app-builder, the team bet that in the agent era the computer itself — what users open, watch and delegate to — is the product.

The bet is still being tested: rivals including Perplexity announced similar "agent-native" computer visions in the same period, and HappyCapy must now prove retention beyond its launch spike. As of 2026-09-02 it continues to scale, shipping weekly skill-marketplace updates and marketing itself as computing "for the rest of us."

What has to be true

  • Friction wedge: OpenClaw showed what agents could do but left self-hosting, Docker and always-on hardware to the user; HappyCapy removed that barrier with a zero-setup cloud sandbox and a GUI.
  • Audience bet: HappyCapy targeted non-technical creators and knowledge workers, not developers, betting an agent-native computer could be sold as a consumer product rather than a dev tool.
  • Fast capital: $1M ARR in the first month and >$10M in funding from HongShan, Baidu VC and others let the team keep shipping while the category was still being defined.
  • Market timing: launching in February 2026, weeks after the OpenClaw wave crested, gave HappyCapy the visibility to take #1 day and week on Product Hunt.

What can be applied

HappyCapy won by removing OpenClaw's setup barrier and wrapping a developer agent in a GUI for everyone; the launch proved demand, but durability against bigger rivals is unproven.

Aftermath

As of 2026-09-02 HappyCapy is live and scaling: desktop browser product plus iOS app, free/Pro/Max tiers, 30k+ skills, 150+ built-in models and weekly skill-marketplace updates, with over $1M ARR in its first month and over $10M raised from HongShan Seed Fund, Planetree VC, Monolith and Baidu VC. The open question is durability: launch momentum was community-driven, and reviews put far more autonomous platforms like Manus and Devin among the alternatives, so HappyCapy must prove that a visible, interruptible "agent computer" retains paying non-technical users.

Sources

spotted an error? The archive wants to know.

Your turn

You just read one. Describe what you are building, and see who is betting on the same thing.

Free account · 3 free questions · no card

Related cases