Skip to content

happycapy Review

Happycapy is a browser-based "agent-native computer": AI agents powered by Claude Code work in a cloud sandbox with skills, scheduled automations and a choice of models. Free plan, paid from $20 a month.

Free plan · from $20/mo

Stork's verdict on happycapy

Happycapy turns Claude Code into a hosted agent computer with skills and schedules; the free plan is a fair test, but track credit use before paying annually.

happycapy reviewed by Stork AI · stork.ai/en/happycapy

720p. 480p is half that. Minimum 5 cents a video.
Ten minutes of talking video costs $2.
We run open models on our own GPUs. When our cost per minute falls, so does the price.
Make a photo talkDevelopers: there is an API

What Happycapy does

Happycapy (happycapy.ai) calls itself "Your agent-native computer": a workspace that runs in your browser where AI agents browse the web, work with files and run code inside an isolated cloud sandbox. Its docs say it is "powered by Claude Code" and routes each task to a suitable built-in model, with no installs, terminal setup or API keys. The FAQ names model families including Claude, MiniMax, DeepSeek and Kimi, with media tasks routed to models such as GPT Image 2, Gemini or Seedream; supported models can run at thinking effort levels from Low to Max, or Adaptive. Product Hunt describes it as built by the original Trickle team.

Work is organised in Projects (with files and a CLAUDE.md for instructions). A Skill Store adds ready-made skills for websites, deep research, image and video generation, games and slides, and you can write your own. Automations (in beta) run a prompt on a daily or interval schedule, and tasks keep running when your computer is off. Connections to outside systems go through MCP servers (the docs give Gmail, Google Calendar, Notion, Slack, GitHub, Linear and Postgres as examples), custom secrets and CLI tools; Connect My Mac lets an agent use local files, commands and apps on a Mac. Capymail gives agents an email inbox on paid plans. It runs in the browser and, per the FAQ, as an iOS or Android app.

Happycapy pricing

PlanMonthly priceAnnual billing (per month)Credits per monthSandboxAlso included
Free$0-250BasicClaude model access, basic models via skills, custom skills
Pro$20/month$17/month2,0002 cores, 4GB RAM, 50GB storageFrontier models, 3 automations, Capymail (200 emails)
Plus$50/month$42/month5,0002 cores, 4GB RAM, 50GB storage5 automations, 2,000 emails
Max$200/month$167/month22,0004 cores, 8GB RAM, 200GB storage10 automations, 5,000 emails, early iOS app access, agent teams (research preview), priority human support
TeamCustomCustomNot statedNot statedContact via "Get Team"

Cost per 100 credits at monthly prices: Pro $20 / 20 = $1.00, Plus $50 / 50 = $1.00, Max $200 / 220 = $0.91. On annual billing (advertised as "2 months free"): Pro $17 / 20 = $0.85, Plus $42 / 50 = $0.84, Max $167 / 220 = $0.76. The page shows annual prices per month and does not say whether the year is charged up front.

Credits are spent on model calls, media generation, long-running tasks and Automations. Happycapy does not publish how many credits a given model or task uses; its docs say reasoning models, higher thinking effort, multi-step research and coding, and image or video skills use more. Usage by model is shown under Settings, Billing. The pricing page and docs do not mention credit top-up packs, rollover or refunds.

Prices checked on the vendor's pricing page on November 10, 2026. See Happycapy pricing.

Key features

What Happycapy's site and docs describe:

  • Isolated cloud sandbox where agents browse, edit files and run code
  • Built on Claude Code, with no API keys to manage
  • Skill Store for websites, research, images, video, games and slides, plus custom skills
  • Automations (beta): scheduled prompts that run daily or on an interval
  • MCP servers, custom secrets and CLI tools for connecting outside services
  • Connect My Mac to work with local files, commands and apps
  • Capymail: an email inbox for agents on paid plans
  • Step-by-step progress view with the option to stop an agent at any time

Who Happycapy is best for

Happycapy suits people who want a Claude Code style agent without setting up a terminal: founders, operators and analysts who hand off research, file processing, small sites or reports, and want them to keep running on a schedule in the cloud. The free plan's 250 credits are enough to try it; regular use needs Pro or higher. Because credit costs per task are not published, start monthly and watch the usage page before moving to annual billing.

Developers comfortable in a terminal may prefer Claude Code directly. For a general autonomous agent with a consumer focus, compare Manus and Genspark; for a canvas-style agent workspace, see Flowith.

What users say

Independent coverage is limited so far. On Product Hunt, Happycapy holds a 4.7 out of 5 rating from 38 reviews at the time of checking, with reviewers praising how autonomous and workflow-oriented it feels and using it for research, reminders, deployments and GitHub or Google Calendar workflows. Criticisms on the same page include no visibility into intermediate steps on long tasks, a remote browser view one user called too laggy to be useful, and automations that did not run while the sandbox was sleeping. Some older reviews there refer to the product by its earlier Trickle name. No larger review base (G2, Trustpilot) was found.

Pros

  • Agent environment in the browser with no installs or API keys
  • Usable free plan with 250 credits a month
  • Scheduled automations keep running when your computer is off
  • Choice of model families and thinking effort levels
  • Extends with MCP servers, CLI tools and a local Mac connection

Cons

  • Credit cost per task or model is not published
  • Automations are still in beta
  • No credit top-ups, rollover or refund terms on the pricing page
  • Small independent review base so far

Frequently asked questions

What is Happycapy?

Happycapy is a browser-based "agent-native computer" where AI agents, powered by Claude Code, work in a cloud sandbox: browsing, editing files, running code and running scheduled automations. It adds a Skill Store, MCP connections and an agent email inbox.

How much does Happycapy cost?

Free is $0 with 250 credits a month. Pro is $20 a month (2,000 credits), Plus $50 (5,000 credits) and Max $200 (22,000 credits); annual billing shows $17, $42 and $167 a month. Team pricing is on request. Prices checked on the vendor's pricing page on November 10, 2026.

Is Happycapy free?

Yes, there is a free plan with 250 credits a month, a basic sandbox, Claude model access and custom skills. Frontier models, automations and Capymail need a paid plan.

How many credits does a Happycapy task use?

Happycapy does not publish per-task or per-model credit costs. Its docs say higher-capability models, higher thinking effort, multi-step research or coding, and image or video skills use more credits; you can see usage by model in Settings, Billing.

Does Happycapy need an API key?

No. Happycapy says frontier models are included in the subscription with no API keys required. You may still need tokens or logins for third-party services you connect through MCP.

Is Happycapy the same as Claude Code?

Happycapy's docs say it is powered by Claude Code, but it runs in a hosted cloud sandbox in your browser with its own skills, automations, models from several providers and a credit-based plan, rather than in your local terminal.

Can Happycapy work with files on my computer?

By default it only uses files you upload or add to a Project in its cloud sandbox. The Connect My Mac feature lets it read and save local files, run local commands and use desktop apps on a connected Mac.

Compare Alternatives

happycapy

Free plan · from $20/mo

Enjoying this? Get one like it in your inbox each morning.

one email a day · unsubscribe in two clicks · no third-party tracking