Capy

Overview

Capy is a cloud coding platform where you hand off work to AI agents and get back pull requests. Start a thread from the web app, Slack, Linear, or the API. Capy sets up a sandboxed VM with your repo, makes the change, opens a PR, and handles CI and review feedback. You can run many agents at once, schedule automations, and pick from models like Claude, GPT, Gemini, and Grok. It suits engineering teams with large backlogs, migrations, and review queues.

Key Features

Parallel cloud agents: Run many threads at once, each in its own isolated cloud machine. Subagents can split large jobs like migrations and refactors across separate or shared VMs.
Captain, Build, and Review agents: The Captain agent plans the work, Build agents write and test the code, and Review agents post findings on your pull requests.
Automations: Trigger tasks on a cron schedule or from platform events, such as triaging bugs posted in a Slack channel or responding to GitHub comments.
Native PR review UI: Create and review pull requests inside Capy, and split, filter, and sort agent sessions in the web UI.
Fast VMs with snapshots: Machines support sandbox snapshotting, setup scripts, secrets, and Docker-in-Docker, so agents can actually run your code.
Flexible AI model access: Choose from supported models across providers such as Anthropic, OpenAI, Google, xAI, and others. Connect your own API keys or compatible subscriptions, including Codex, GitHub Copilot, and SuperGrok, where supported by your plan.
Team integrations: Connect GitHub, Slack, Linear, Sentry, Vercel, Tailscale, and MCP servers so tasks start and report back where your team already works.
REST API: Start threads, send messages, interrupt, archive, list tasks, and launch reviews programmatically.

Pros

Unlimited members on every plan: There are no per-seat fees, so you can add your whole team without the bill growing per person.
Built for throughput: Running many agents in parallel suits backlog cleanup, large migrations, and bulk refactors better than one assistant working through tasks one at a time.
Works where your team works: You can start tasks from Slack, Linear, the web app, or the API, and progress streams back to those tools.
Model flexibility: You aren't locked into one AI provider, and you can bring your own keys or existing subscriptions to control costs.
Security credentials: Capy is SOC 2 Type II certified, with encrypted, isolated execution environments and a public trust center.
Low-cost trial: Try Capy Lite for $1 for seven days with a verified payment card before the subscription renews at $20 per month.

Cons

No permanent free plan: Capy offers a $1 seven-day introductory trial that requires a verified payment card, then renews at $20 per month on the Lite plan. Users looking for a free option for occasional testing may prefer other tools.
Usage-based credit limits: AI model calls and cloud VM compute use monthly credits. Unused credits do not roll over, and additional usage requires a top-up or plan upgrade. Heavy workloads can cost more than the base subscription.
Cloud-based coding workflow: Capy focuses on cloud-based coding agents, task threads, and pull requests rather than traditional inline code autocomplete. Developers who prefer real-time coding suggestions inside their editor may prefer an IDE-focused assistant.

Use Cases

Backlog and bug triage: Set up an automation that picks up bug reports from a Slack channel or Sentry, investigates them, and opens fix PRs. Your team reviews finished work instead of starting from scratch.
Large code migrations: Split a framework upgrade or API migration across parallel subagents. Each one works in its own VM and branch, so a multi-week effort becomes a set of reviewable PRs.
Automated PR review: Run Capy's review agent on every pull request to catch issues and do visual QA before a human reviewer looks at it.
Ticket-to-PR from Linear: Delegate a Linear issue to Capy and follow its progress in Linear. It returns a pull request after CI has run.
End-to-end testing: Use agents with browser and computer use to run through app flows and produce annotated videos that show what was tested.
Documentation upkeep: Schedule recurring tasks that update docs to match recent code changes, so documentation doesn't drift behind the codebase.

Frequently Asked Questions

What is Capy?

Capy is a cloud coding agent platform. You describe a change in a thread, and Capy sets up a machine with your repository, writes the code, opens a pull request, and handles CI and review feedback.

How much does Capy cost?

Lite: $20/month or $16/month with annual billing, including $20 in monthly credits.
Pro: $100/month or $80/month with annual billing, including $105 in monthly credits.
Max: $200 to $1,000/month, or $160 to $800/month with annual billing, with 10% bonus credits.
Enterprise: Custom pricing with tailored credit allowances and additional team controls.

Is there a free trial?

Yes. The first week of the monthly Lite plan costs $1 with a verified card, then it renews at $20/month. It is limited to one trial per card and per organization.

Do I pay per seat?

No. Every plan includes unlimited members without per-seat fees. Your subscription cost depends on the plan, included credits, and available cloud computing resources.

What do credits cover, and do they roll over?

Credits cover AI model costs and VM compute time. Unused credits don't roll over. You can add balance at 1:1 value (minimum $5), turn on auto-reload, or upgrade your plan.

Which AI models does Capy support?

Capy supports models from major providers, including Anthropic's Claude, OpenAI's GPT, and Google's Gemini, along with other supported models. All plans support bringing your own API keys and gateways. Compatible ChatGPT (Codex), GitHub Copilot, and SuperGrok subscriptions can also be connected.

Which tools does Capy integrate with?

Capy integrates with GitHub, Slack, Linear, Vercel, Tailscale, and MCP-compatible servers. It also provides API access on every plan, allowing developers to start coding tasks, launch reviews, and retrieve results programmatically.

Can Capy create pull requests automatically?

Yes. Capy can make code changes, create branches, and open pull requests for review. Agents can also run tests and respond to review feedback. Developers retain control over whether changes are approved and merged.