/
PreiseBlogAnmeldenKostenlos starten
Zurück zum Blog
Updates

Preview Environments, Pipeline Dashboard, and OpenAI in the Coding Runner

New live previews directly in tickets, a cross-provider pipeline dashboard, OpenAI models in the coding runner, project templates, and new automation triggers for pull request status.

Spedy Team5 min readAuf Deutsch lesen
Preview Environments, Pipeline Dashboard, and OpenAI in the Coding Runner
#preview#pipelines#runner#openai#automations#billing#project-templates

Five major updates in this release: live preview environments directly in tickets, a new pipeline dashboard for all CI/CD providers, OpenAI models in the coding runner with per-provider authentication, project templates for quick project setup, and new automation triggers for pull request status.


Preview Environments: the running app right inside the ticket

You can now launch a live preview of your project directly in a ticket — no local dev environment, no manual deployment.

How it works

  1. Open a ticket whose board is connected to a repository
  2. Click Open preview
  3. Spedy starts a Docker container with your project and shows the running app in an embedded browser

The preview opens in a split view: the app on the left, ticket details on the right. You can navigate within the app — the address bar shows the current path and lets you jump to specific routes.

What you can do in the preview

  • Navigate: Click through the app or type a path in the address bar
  • Annotate: Click a UI element to leave a comment — the annotation is saved to the ticket
  • Use the AI agent: In the Agent tab (right sidebar), describe what you want to change. The agent sees the current preview context, makes code changes, and the preview auto-refreshes
  • Add screenshots: Paste or upload screenshots so the agent has visual context

Multiple views per project

If your project has multiple entry points (e.g. /admin and /storefront), define them in a .spedy/preview.yml. In the preview, switch between views via a dropdown.

Configuring the preview stack

Under Project Settings → Agents, configure:

  • Repository: the GitHub repo used for previews
  • Preview stack: which Docker stack powers the preview (Shopware, Magento, Next.js, or custom)
  • Snapshots: optional database dumps so previews start with test data

Project Templates: new projects in one click

Under Settings → Project Templates you'll find preconfigured repository blueprints. When creating a new project, pick a template — Spedy clones the repository and sets up the project automatically.

Built-in templates

  • Shopware 6 (dockware) — e-commerce with preconfigured preview stack
  • Magento 2 (Mage-OS) — e-commerce alternative
  • Next.js — frontend projects

Adding custom templates

  1. Go to Settings → Project Templates
  2. Click Add template
  3. Enter the name and GitHub repository in owner/repo format
  4. Optional: add a description

The repository must be a GitHub template repository. Built-in templates can't be deleted; custom ones can.


Pipeline Dashboard: all CI/CD runs at a glance

Under Pipelines in the main navigation you'll find a new, cross-provider overview of all CI/CD pipeline runs. Whether your team uses GitHub Actions, GitLab CI, or Bitbucket Pipelines — it's all in one place.

What the dashboard shows

  • Status filters: Needs Attention, Running, Failed, Success — or all
  • Provider filter: GitHub, GitLab, Bitbucket — individually or combined
  • Per run: PR author, branch, repository, status, duration, and count of failed/passed steps
  • Real-time: New runs and status changes appear automatically via WebSocket

Click a run to see individual pipeline steps and their details. Pipelines is a feature you can enable in your board's Feature Settings.


OpenAI models in the Coding Runner

The coding runner now supports OpenAI models alongside Claude. Pick the model — the provider follows automatically.

Available models

Anthropic (Claude Code):

  • Haiku 4.5 — fast, affordable
  • Sonnet 4.6 — balanced
  • Opus 4.8 — strongest, new default

OpenAI (Codex):

  • GPT-5 Codex — optimized for code
  • GPT-5 — general purpose

Model selection is in your Runner Team Settings under pipeline stages and agent configurations. You can choose a different model per stage and per agent.

Per-provider authentication in the setup wizard

The runner setup wizard now separates AI backend configuration by provider:

  • Claude: Off / API key (via vault) / Subscription (local Claude Max login on the runner)
  • OpenAI: Off / API key (via vault) / Subscription (local ChatGPT login on the runner)

Both providers can be active at the same time. The runner routes jobs to the right backend automatically, based on the selected model.


GitLab repositories in the Coding Runner

The coding runner can now clone and work with GitLab repositories. When you connect a GitLab repo to your board, the runner detects the clone URL and sets itself up accordingly — including self-hosted instances.


New automation triggers: PR pipeline and review status

The automations engine now understands pull request status as a condition. You can build rules that react to pipeline status (Success, Failed, Running) and review status (Approved, Changes Requested, Pending).

Example

When the PR status changes, and the pipeline is Success, and the review is Approved, then move the ticket to Done.

Two new trigger events are available:

  • PR Status Changed — fires on any change to pipeline or review status
  • PR Merged — fires when a pull request is merged

Automatic triage for failed pipelines

Organizations can now enable automatic triage for failed CI pipelines. When turned on, Spedy periodically scans open pull requests with failed pipelines and automatically starts a coding agent job that attempts to fix the failure.

The feature is opt-in and respects per-PR cooldowns and retry limits to prevent duplicate jobs.


Billing: card on Pro trial, new invoices

Pro trial with card

The Pro trial (30 days free) now requires a payment method at sign-up. The flow:

  1. Enter your billing address
  2. Checkout via Mollie — a small verification amount is charged and immediately refunded
  3. Trial starts (30 days, €0)
  4. At expiry: automatic conversion to a paid subscription

You can cancel the trial at any time. Access stays until the end of the 30 days.

New invoices

Invoices are now §14-UStG-compliant PDFs with full VAT breakdown. Under Settings → Billing you can see all invoices with status (Paid, Issued, Voided) and download them as PDF.

For EU B2B customers with a valid VAT ID, the reverse-charge mechanism applies. Your VAT ID is automatically validated via VIES when you save your billing address.

Prorated billing on seat changes

When you add a new team member mid-cycle, the prorated amount is charged immediately — not deferred to the next invoice.

Frequently asked questions

Quick answers to the most common questions about this topic.

What are Preview Environments?
Preview Environments spin up an isolated Docker container with your project right inside a ticket. You see the running app in an embedded browser, can navigate, annotate UI elements, and let the AI agent make code changes — without touching your local dev environment.
Where do I find the Pipeline Dashboard?
In the main navigation under Pipelines. It shows all CI/CD runs from your connected repositories — GitHub, GitLab, or Bitbucket — in a single view.
Can I use OpenAI models in the runner now?
Yes. In your runner team you can choose between Claude models (Haiku, Sonnet, Opus) and OpenAI models (GPT-5, GPT-5 Codex) per pipeline stage and per agent. The provider follows the model automatically.
What are Project Templates?
Templates are preconfigured repository blueprints for creating new projects in one click. Spedy ships templates for Shopware, Magento, and Next.js — you can also add your own GitHub template repos.
What changes in billing?
Pro trials now require a card at sign-up. The 30 days are free, then billing starts automatically. Invoices are now §14-UStG-compliant PDFs with full VAT breakdown.
Preview Environments, Pipeline Dashboard, and OpenAI in the Coding Runner | Spedy Blog