Add-ons

Add-ons are optional features you can attach to your plan to extend its capabilities. They are independent of your base plan and can be added or removed at any time from the Previewops dashboard.

Advanced Perf

Price: $29/month
Available on: Premium BYOC, Pro, Enterprise, Custom

The Advanced Perf add-on upgrades the /perf-previewops performance testing experience:

What it enables

• 60-minute tests: the base duration cap of 30 min is doubled to 60 min, letting you run sustained load tests against your preview environments.
• --set-baseline: the --set-baseline flag becomes available on /perf-previewops, saving the current run as the reference point for all future comparisons on the same repository.
• Baseline comparison: every subsequent perf run for the repository is automatically compared against the saved baseline, and the PR comment shows deltas for p50/p95/p99 latency, throughput, and error rate.

How to activate

1. Log in to the Previewops dashboard.
2. Click Add-ons in the left sidebar.
3. Click Subscribe next to Advanced Perf.
4. Complete the Stripe checkout. The add-on activates immediately.

How to cancel

Cancel the Advanced Perf subscription from the Stripe customer portal (accessible from the dashboard). On cancellation, --set-baseline and baseline comparison are disabled and the duration cap reverts to 30 min. Existing baseline data is retained.

Custom LLM

Price: Free
Available on: All plans (including Free BYOC)

The Custom LLM add-on lets your organisation use your own LLM API key for AI-powered analysis across both /qa-previewops and /perf-previewops.

This add-on is required for Premium BYOC to receive any AI features. Pro, Enterprise, and Custom plans include AI out of the box via the shared operator key; this add-on lets those plans opt in to their own key instead.

This is useful if you want:

• AI features on the Premium BYOC plan (which does not include a shared platform LLM key).
• Full control over which model and provider is used for analysis.
• Your code diffs to remain within your own LLM account.
• Access to AI features on the Free plan (which also has no shared LLM key, but note that QA and perf are blocked on Free).

Supported providers

You choose the model when you save your credentials — the model dropdown updates automatically to show only the models for the selected provider, with the recommended one preselected.

The recommendations are the models we have found produce the most useful code review, not the cheapest. Benchmarking them against each other on a large multi-language pull request, claude-opus-5 found nine issues to claude-sonnet-5's six, and the two disagreed by two severity levels on the most serious of them — a review you can't trust on severity is worth less than the token saving. Since you supply the key, the choice and the cost are yours: a cheaper model in the same dropdown is one click away.

Review depth does vary between providers as well as between models, and more so on large diffs than small ones. We won't publish a league table — it would be out of date within a month — but the recommended model for each provider is the one we have actually measured on a large multi-language pull request, so it is the safest starting point. If a review reads as thinner than you expected, trying another provider's recommended model is a cheap experiment.

For Gemini specifically the recommendation stays on the Flash tier. The only current-generation Gemini Pro model is a -preview endpoint, and a preview endpoint is a poor default: Gemini takes the model name in the request URL, so if the endpoint is withdrawn every run fails and the credential is marked inactive. Both gemini-3.1-pro-preview and the Flash-Lite tier remain selectable if you'd rather make that trade yourself.

How to activate

The Custom LLM add-on is free and requires no Stripe checkout. Activate it directly from the dashboard:

1. Log in to the Previewops dashboard.

2. Click Add-ons in the left sidebar.

3. Click Activate next to Custom LLM.

4. Add your LLM credentials in the Credentials section: enter your LLM_API_KEY, select LLM_PROVIDER (openai, anthropic, or gemini). Optionally expand Advanced options to choose a specific LLM_MODEL — defaults to the recommended model for the provider if omitted.

   Advanced options also expose per-request diff budget controls (LLM_MAX_TOTAL_PATCH_CHARS, LLM_QA_MAX_FILES_FOR_PROMPT, LLM_QA_FALLBACK_TOTAL_PATCH_CHARS, LLM_QA_MAX_OUTPUT_TOKENS) for tuning token usage on repositories with very large PRs. See the Credentials guide for details.

Preview DB

Price: $39/month flat + $2.00/branch above 30 per month Available on: Premium BYOC, Pro, Enterprise, Custom

The Preview DB add-on provisions a fresh, isolated Postgres database branch for every pull request — automatically. It uses database branching under the hood and requires no changes to .previewops.yaml.

What it enables

• Per-PR database isolation: each PR gets its own database branch forked from your organisation's staging database snapshot. Data changes in one PR's preview do not affect any other PR.
• Automatic DATABASE_URL injection: both DATABASE_URL and PREVIEW_DB_URL are injected into the preview container at deploy time — no env: configuration needed.
• Automatic cleanup: the database branch is deleted when the PR is closed or the preview's TTL expires.
• Nightly resync: the parent branch (source for all PR forks) is automatically refreshed from your staging database URL every night.

Branch cap and overage

Your plan includes 30 Preview DB branches per month. Branches created above that cap are billed at $2.00 per branch via a Stripe metered component. The monthly counter resets on the 1st of each month.

How to activate

1. Log in to the Previewops dashboard.
2. Click Add-ons in the left sidebar.
3. Click Subscribe next to Preview DB.
4. Complete the Stripe checkout. The add-on activates immediately.

How to set up your staging URL

After activating the add-on, store your staging database URL as a credential:

1. Click Credentials in the left sidebar.
2. Set Provider key to preview-db.
3. Add the key PREVIEW_DB_STAGING_URL with your staging Postgres connection string as the value.
4. Click Save.

Previewops will sync this database into the preview parent branch on the next nightly job (or immediately if you trigger a manual resync from the dashboard).

The staging URL must be a publicly reachable Postgres connection string. Private-network-only hosts are not supported.

Keeping the parent branch in sync

The parent branch is automatically resynced every night. You can also trigger a resync on demand from the dashboard:

1. Click Add-ons in the left sidebar and click Preview DB → Resync now.

How to cancel

Cancel the Preview DB subscription from the Stripe customer portal (accessible from the dashboard). On cancellation, new PR deploys no longer receive a database branch and DATABASE_URL / PREVIEW_DB_URL are no longer injected. Existing previews still running at cancellation time retain their branches until they expire or are deleted.

Plan eligibility summary