const agent = await client.create({
  model: "claude-sonnet-4-5",
  tools: ["read", "write", "bash"],
});

def fix_tests(repo: str):
    agent = Agent(model="sonnet")
    result = agent.run(
        f"Fix failing tests in {repo}"
    )
    return result.pr_url

impl Agent {
    pub async fn run(&self) -> Result<PR> {
        let diff = self.generate_changes()?;
        self.create_pull_request(diff).await
    }
}

{
  "status": "complete",
  "changes": 14,
  "pr_url": "github.com/org/repo/pull/42"
}

func (a *Agent) Schedule(cron string) error {
    return a.scheduler.Add(cron, a.Run)
}

diff --git a/src/auth.ts
@@ -12,4 +12,6 @@
-  return input;
+  const sanitized = validate(input);
+  return sanitized;

schedule:
  cron: "0 4 * * 1-5"
  repo: your-org/app
  auto_pr: true

$ curl -X POST /api/v1/agents \
  -d '{"prompt": "Update deps"}'
> 201 Created {"run_id": "r_8f3k"}

SELECT r.id, r.status, r.pr_url
FROM agent_runs r
WHERE r.completed_at > NOW() - '1d'
ORDER BY r.created_at DESC;

Run coding agents
with an API call

Run OpenCode against your GitHub repo with a simple API call. No infrastructure to manage.

Documentation

terminal

curl -X POST https://codecloud.dev/api/v1/agents \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "repo": "your-org/your-repo",
    "prompt": "Fix the failing tests in src/utils",
    "model": "claude-sonnet-4-5",
    "webhook_url": "https://your-app.com/webhook"
  }'

How it works

Three steps to automated PRs

Connect your repo

One-click GitHub app install. Select which repos to grant access to.

Send a prompt

REST API, Linear assignment, or any automation tool you already use.

Get a PR

Review the diff, leave comments to iterate, then merge when ready.

Enterprise-grade security

Your code and credentials are protected at every step

Encrypted at rest

API keys protected with AES-256-GCM encryption

Short-lived tokens

GitHub access uses encrypted, temporary credentials

Isolated sandboxes

Each agent runs in its own secure, isolated environment

No data retention

Repository contents deleted permanently after each run

Learn more in our Privacy Policy

Built for automation

Integrate however fits your workflow

API

REST API

Full control via HTTP. Start runs, poll status, receive webhooks. Add AI coding capabilities to any product or service.

View docs

n8n / Zapier / Make

Build no-code workflows that write code. Trigger agents from forms, emails, Slack messages, or any other event.

See n8n guide

Linear

Native

Assign issues to @codecloud. Watch the agent think in comments. Talk back to course-correct.

Learn more

CI/CD & GitHub Actions

Trigger agents from your pipeline. Auto-fix failing tests, generate docs on merge, or run code review bots on every PR.

View docs

Any modelbring your own API keysZero infrastructurewe handle computeAuto PRsready to review and merge

Put maintenance on autopilot

Schedule agents to run daily, weekly, or on any cron cadence. They work while your team sleeps and open PRs for review in the morning.

Dependency updates

Go beyond Renovate — update packages and fix the breaking changes automatically.

Tech debt cleanup

Remove dead code, fix linter warnings, migrate deprecated APIs — a few files at a time.

Security audits

Run regular scans and let the agent patch known vulnerabilities before they become issues.

Every weekday at 4 AM

POST /api/v1/schedules
{
  "type": "recurring",
  "cron": "0 4 * * 1-5",
  "repo": "your-org/your-repo",
  "prompt": "Update npm dependencies and fix any breaking changes",
  "auto_create_pr": true
}

Learn more about scheduled runs

More than a coding agent

Give agents custom tools, run them without a repo, or use plan mode to review before any code changes

Custom tools

Pass HTTP endpoints as tools so agents can call your internal APIs, query databases, or check feature flags while they work.

See tool use guide

Blank workspace

Run agents without a repo for non-code tasks — data transforms, report generation, or multi-system orchestration via tools.

Learn more

Plan mode

Generate implementation plans grounded in your actual codebase before writing any code. Great for estimating and ticket validation.

See plan mode guide

Autonomous iteration

Agents keep working until tests pass, not just until they think they're done. Set success criteria and let them loop.

View docs

Any model — bring your own API keysZero infrastructure — we handle computeAuto PRs — ready to review and merge

Simple pricing

Codecloud is just a fixed monthly price on top of your usual LLM API usage

No plans available. Please configure plans in your Clerk dashboard.

Frequently asked questions

Everything you need to know about Codecloud

Have more questions? Contact support

Ready to get started?

Connect your repo and start running agents today.

Run coding agentswith an API call

Three steps to automated PRs

Connect your repo

Send a prompt

Get a PR

Enterprise-grade security

Encrypted at rest

Short-lived tokens

Isolated sandboxes

No data retention

Built for automation

REST API

n8n / Zapier / Make

Linear

CI/CD & GitHub Actions

Put maintenance on autopilot

Dependency updates

Tech debt cleanup

Security audits

More than a coding agent

Custom tools

Blank workspace

Plan mode

Autonomous iteration

Simple pricing

Frequently asked questions

Ready to get started?

Run coding agents
with an API call