A clearer view of AI.
THE INDEPENDENT AI.

CODING AGENT / DEV TOOL · WORTH A LOOK

Coldtea

macOS agentic development environment that runs your coding agents, regression-tests every PR with QA agents and watches production, filing what…

Visit coldtea.ai ↗@ColdteaAI on X ↗

What Coldtea does

macOS agentic development environment that runs your coding agents, regression-tests every PR with QA agents and watches production, filing what breaks as tasks. A desktop IDE built around the terminal: runs any CLI coding agent (Claude Code, Codex, Gemini CLI, OpenCode) in parallel git worktrees with split panes, a built-in editor, diff annotation and agent teams. Tasks sync two-way with Linear, Jira and GitHub Issues and keep each run's plan and logs. QA agents open each PR's preview URL and click through the app like a user (tests written in plain English, on web, desktop, iOS and Android), then re-run on production. Monitoring agents read Sentry, PostHog, Datadog, Grafana, New Relic, Vercel and logs, and file bugs back as tasks. Optional cloud background agents (off by default; code stays local). The same team makes PR Lens (prlens.dev), an open-source (MIT) tool that draws architecture and data-flow diagrams inside pull requests, free for open source.

Good for

  • Automatically regression-test client apps by pointing QA agents at a preview URL
  • Run several coding agents in parallel on one task without collisions
  • Turn production errors from Sentry/PostHog into ready-to-fix tasks
  • Make AI-written PRs easier to review with PR Lens diagrams

Try this in 10 minutes

  1. Automatically regression-test client apps by pointing QA agents at a preview URL. Use the free plan: pick one real example from your own business and try just that.
  2. Run several coding agents in parallel on one task without collisions. Put the result next to how you do it today. Is it faster, cheaper or better?
  3. Turn production errors from Sentry/PostHog into ready-to-fix tasks. Note what you had to fix by hand before you'd show it to a customer.

Mini-tasks based on common small-business uses for this tool.

Honest take

Some claims not independently checked

  • More developer-focused than most of the directory, but the URL-only QA agent could be useful for testing apps built with no-code/vibe-coding tools; compare with other agentic IDEs and AI QA tools.
  • Testimonials on the site are vendor-selected.
  • Founder/MD per imprint: Emmanuel Ohanekwu (@OhansEmmanuel).

What we couldn’t verify

  • Site claims ISO 27001, SOC 2 Type 2 in progress.

Pricing

Freemium: terminal free forever (no seat count); monitoring free with your own keys; agentic testing 2,000 free credits/month, extra credits paid (price Unknown). PR Lens free for open source; paid PR Lens pricing Unknown.

Prices as listed by the maker when we last checked (2026-10-18); usually in US dollars. Check the website for current prices.

Latest update

  • : Every pull request now ships with an interactive walkthrough (beyond Mermaid diagrams).

Is Coldtea worth it? Quick answers

Is Coldtea worth it?

Coldtea is worth a look if it matches a job you already do (Coding agent / dev tool). More developer-focused than most of the directory, but the URL-only QA agent could be useful for testing apps built with no-code/vibe-coding tools; compare with other agentic IDEs and AI QA tools. There is a free tier, so you can test it before paying. Some of the maker's claims have not been independently checked.

How much does Coldtea cost?

Freemium: terminal free forever (no seat count); monitoring free with your own keys; agentic testing 2,000 free credits/month, extra credits paid (price Unknown). PR Lens free for open source; paid PR Lens pricing Unknown.

What is Coldtea used for?

macOS agentic development environment that runs your coding agents, regression-tests every PR with QA agents and watches production, filing what… Typical uses: automatically regression-test client apps by pointing QA agents at a preview URL; run several coding agents in parallel on one task without collisions; turn production errors from Sentry/PostHog into ready-to-fix tasks.