docs(billing): define agent hours for platform credits and refresh stale language - #672
docs(billing): define agent hours for platform credits and refresh stale language#672hongyi-chen wants to merge 5 commits into
Conversation
…ale language Platform credits are calculated from agent hours; add the definition to the platform credits page (intro, description, and measurement section) and echo it on the credits, pricing FAQs, and enterprise billing pages. De-dupe the intro vs the three-bucket section, cut AI-isms, and update the pricing FAQs preview-period language now that platform-credit billing is live. Co-Authored-By: Warp <agent@warp.dev>
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
|
This PR was generated with Warp. Comment |
|
Your GitHub account is not connected to Warp. Connect it here. |
|
I'm starting a first review of this pull request. You can view the conversation on Warp. I completed the review and no human review was requested for this pull request. Comment Powered by Oz |
There was a problem hiding this comment.
Overview
This PR updates billing documentation to define agent hours for platform credits, refresh platform-credit billing language after the preview period, and tighten related prose across the billing pages. I reviewed the changed docs against the attached diff, docs style guidance, security checklist, and the provided spec context; no approved implementation spec was available for drift checks.
Concerns
- No blocking concerns found.
- Comment audit: the diff does not add code comments or documentation comments that violate repository conventions.
- Test audit: the diff does not add or change tests.
Verdict
Found: 0 critical, 0 important, 0 suggestions
Approve
Comment /warp-agent-review on this pull request to retrigger a review (up to 3 times on the same pull request).
Powered by Oz
There was a problem hiding this comment.
Overview
Refreshes the platform-credits docs to define "agent hours," updates stale preview-period language now that billing is live, and trims AI-ism prose across four billing/credits pages. No correctness issues found; recommend merging as-is.
Concerns
- Four pre-existing instances of "Warp's platform infrastructure" remain in the very pages this PR is refreshing to "platform layer" terminology (
pricing-faqs.mdx:326,336andenterprise/support-and-resources/billing.mdx:46,91, all outside the diff). Not introduced by this change, but worth a follow-up pass for full terminology consistency on these pages.
Verdict
Checks: build pass, internal links 3968/3968 pass (0 broken), style_lint --changed clean (all 10 reported findings are pre-existing/false-positive, none on changed lines), visual proof n/a (docs-only prose change)
Found: 0 critical, 0 important, 1 suggestion, 0 nits
Responding as revops-factory: Open session · View factory task
…tories in intro The platform credits meta description names both charged scopes again (cloud agent runs and local runs with customer-supplied inference), and the intro now mentions Warp Factories runs up front instead of only in the eligibility bullet. Co-Authored-By: Warp <agent@warp.dev>
Platform-credit billing on self-serve plans started July 1, 2026, so the "When do platform credits start being charged" question no longer needs its own FAQ entry. No pages link to its anchor. Co-Authored-By: Warp <agent@warp.dev>
- Correct the agent-hour boundary: the clock starts when the agent starts in its environment (including environment setup), not after setup. Verified against the sandbox driver, which reports IN_PROGRESS before environment setup and MCP startup. - Add the self-hosted-workers case to the compute credits section: those cloud runs never consume compute credits but still consume platform credits (server rejects compute charges for self-hosted runs). - Replace 'don't consume Warp credits' with 'don't consume AI credits' in the pricing FAQs where the same answer goes on to say platform credits apply; platform credits are Warp credits. - Fix the add-on credits page's inverted claim that add-on credits 'draw from the same pool as platform credits' (add-on credits are a source, platform credits a usage type). - Normalize the credits page heading hierarchy (H2/H3, no bold in headings), replace the versioned model list with a non-drifting statement, fix 'Warp's Agent' terminology, and broaden the integrations list. - Standardize the agent-hour phrasing across all billing pages and trim restated cause-and-effect on the add-on credits page. Co-Authored-By: Warp <agent@warp.dev>
| --- | ||
|
|
||
| This page covers general information about credits, cloud agent costs, and BYOLLM billing for enterprise teams. If you have specific questions about custom pricing, credit allocation, or contract terms, contact your Warp account manager. | ||
| Warp Enterprise uses the same credit-based billing as other plans, with pool sizes, rates, and terms set by your contract. For questions about custom pricing, credit allocation, or contract terms, contact your Warp account manager. |
There was a problem hiding this comment.
this will update with pricing transparency, but this is still accurate today


Summary
Sales needs the platform credits docs to state that platform credits are calculated from agent hours, with a clear definition. The term appeared nowhere in the docs. This PR adds the definition to the platform credits page and echoes it on the three sibling billing pages, refreshes stale preview-period language now that platform-credit billing is live (prod has been in charge mode since July 1, 2026), and de-slops the prose per the AGENTS.md voice rules.
Behavior claims were verified against the warp-server billing code (
logic/ai/ambient_agents/billing/platform.go,specs/REV-1551/TECH.md) and the sandbox driver in the warp client (app/src/ai/agent_sdk/driver.rs): platform credits accrue per minute of wall-time from when the agent starts in its environment (the driver reports IN_PROGRESS before environment setup and MCP startup, so setup is inside the billed window) until the run ends or pauses for input. Queue and sandbox-provisioning time before the agent starts, waiting-on-input time, and idle time aren't billed. Exact rates stay deferred to warp.dev/pricing (per-tier and still tuned server-side).Changes
The plans-and-billing pages live in
src/content/docs/support-and-community/plans-and-billing/.platform-credits.mdx
#the-three-credit-bucketsanchor, which three other pages link to.credits.mdx
maxWidthto the usage-footer screenshot, replaced the filler-opener frontmatter description.pricing-faqs.mdx
add-on-credits.mdx
index.mdx and plans-pricing-refunds.mdx
enterprise/support-and-resources/billing.mdx
Content design plan
Audience and JTBD: A prospect or customer in a sales conversation (and the account team supporting it) who needs to understand what platform credits are and how they're calculated.
Problem: The docs described the metering behavior but never named or defined "agent hours", the term sales uses; the pricing FAQs still described the pre-July-2026 preview period in present tense.
Content type: Update to existing feature/FAQ pages — no new page.
Unverified claims
None — metering behavior (including the agent-hour start boundary), charge eligibility, the self-hosted compute exemption, and billing-start dates were verified against warp-server source (
logic/ai/ambient_agents/billing/platform.go,logic/ai/ambient_agents/finalization.go,model/types/billing.go,config/prod.yaml,specs/REV-1551/TECH.md) and the warp client (app/src/ai/agent_sdk/driver.rs,app/src/settings_view/billing_and_usage/billing_cycle_usage_common.rs).Follow-ups not in this PR
reload-credits.png(add-on credits page) shows an internal test account email and a "5 credits / 99% off" test pack. Worth re-capturing from a clean account.credits.mdxstill has 7 callouts (mostly per-factor "Tip" notes); folding them into body prose is a bigger restructure than this PR.credits.mdxpredates the GitHub, GitLab, Bitbucket, Azure DevOps, and Jira integrations; not verified here.Validation
style_lint --changed: no findings on changed content (remaining findings are pre-existing: page-wide callout density, proper-noun header-case false positives, glossary suggestions for bolded link text).check_for_broken_links --internal-only: 3,966 internal links, 0 broken. Heading-level changes oncredits.mdxdon't alter slugs.npm run build: passes.Plans:
Co-Authored-By: Warp agent@warp.dev