Hand Canopy a repository and a job.

Give Canopy a repository and a task. It plans, edits files and runs the project’s checks in a local or cloud workspace, and you review the result.

Illustration: From task to tested change. Choose a workspace: Attach a folder or select a repository; Review the change: Follow file edits and test runs; Keep what worked: Reuse the workflow as a skill.

The problem

Sound familiar?

“I spend half my day on fixes I already know how to make.”

“Every agent session starts from zero, and I explain the same steps again.”

“I want the agent working while my laptop is closed.”

What it does

From task to tested change.

Attach a local folder in the desktop app, or select a repository and branch for a cloud session. Ask Canopy to trace a bug, propose a plan, edit files and run the project’s checks in its workspace.

Review file changes and command results from the same session. Turn a repeatable fix into a skill, so the next session can follow the same steps, and connect plugins and tools to extend what Canopy can do.

Plan, edit, test

Canopy works the task in the repository.

Local or cloud

Run on your machine or in a cloud runner.

Skills that stick

Reuse what worked in the next session.

Your tools

Connect plugins and MCP servers.

Setup

How it works in Canopy

4 steps, from setup to the work getting done.

  1. Step 1: Choose where it works

    Attach a local folder in the desktop app, or choose a cloud environment with the connected GitHub repository and branch.

  2. Step 2: Set up the workspace

    For a cloud runner, define the setup and start scripts your project needs, plus network access and environment secrets for the job.

  3. Step 3: Describe the job

    Ask Canopy for the fix, the feature or the investigation. It proposes a plan, edits files and runs the checks, with approvals where your settings require them.

  4. Step 4: Review and reuse

    Follow the runner’s state and the command results from the session, review the changes, and keep a repeatable fix as a skill after you approve it.

Sample

Sample conversation, illustrative

A developer hands Canopy a failing test in a cloud session.

Text threadIllustrative
Developer
The invoice date test is failing on the main branch. Find out why and fix it.
Canopy
The parser assumes local time. I propose parsing in UTC, and I will update the test and run the package checks.
Developer
Go ahead.
Canopy
Done. Two files changed and the package checks pass. Review the changes when you are ready.
Developer
Save that as a skill for the next date bug.
Which plan

The plans that carry this.

Suite
Recommended
$45 /user/mo
billed monthly
  • 75M Canopy tokens a month
  • 100 AI resolutions
  • 1,500 outgoing calling minutes (incoming unlimited)
  • 350 text messages (Canada and United States)
  • See everything included
Canopy
$10 /user/mo
billed monthly
Try it

Call the demo line, or try the chat.

Demo line coming soon.

Your next customer is about to call.

Start free, or book a demo.