docs: clarify pricing transparency and usage billing

[hongyi-chen]

What this feature does

Customers can distinguish their usage balance from payments, understand what contributes to run costs, and manage additional usage without confusing dollar billing with existing Enterprise credit contracts. This updates existing documentation only. Production rollout is not independently verified; no feature-specific shipped version or launch date is asserted.

Changes

• Update the canonical usage, purchase, platform-billing, plan, and pricing-FAQ pages; keep detailed answers with their owning pages.
• Align related inference, CLI, Enterprise, team, factory, and troubleshooting guidance, including unattended-run billing and resource-dependent charges.
• Remove obsolete discounts, universal minimum-credit claims, and billing screenshots. Preserve published URLs, legacy anchors, refund policies, and exact public identifiers.
• Align sidebar labels and the billing glossary. Scope: 59 existing documentation pages and three navigation/style-reference files; no product-code changes or new pages.
• Review revision: clarify billing units and billable time, distinguish Free purchases from paid-account controls, centralize cloud usage requirements, and restore the concurrency confirmation placeholder.
• Earlier follow-up: correct current purchase terminology, shorten duplicate FAQ guidance, scope conversion guidance to self-serve purchased packs, and use the preserved monthly-limit anchor in auto-reload instructions.
• Copy revision: restore “Interactive sessions can consume more usage if…” and short explanatory paragraphs in Full Terminal Use, replacing the “keep usage down” tips without changing billing guidance.
• Final review revision: use current user-facing terms, state public API-rate pricing and Free-plan behavior plainly, preserve the legacy monthly-limit anchor, and clarify cloud-run and BYOK billing guidance.
• Docs review follow-up: condition converted-balance and Enterprise pay-as-you-go guidance on the account state and service agreement, and centralize agent API key billing rules.
• Current-head review follow-up: remove unverified rollout, conversion-example, plan-change, and cloud eligibility claims while preserving source-backed conditional guidance.
• Team-page follow-up: cite Free-plan purchased-usage eligibility and restore paid-seat proration when a member rejoins.
• Required-review follow-up: remove display-control availability instructions, condition post-plan-change expiration guidance on the balance remaining available, and scope public API-rate pricing to dollar-billed self-serve plans.
• Previous-head review follow-up: remove uncited CLI unit-switching and web Usage-panel behavior, scope the agent FAQ's API-rate statement, and restore the established QuotaLimit guidance.
• Cross-page editorial follow-up: revise fourteen public pages to distinguish personal and team-managed inference, app and CLI credentials, local and cloud execution, Warp charges and provider bills, and purchase controls and Enterprise consumption limits. Remove backend walkthroughs and speculative setup claims; connect the exhausted-balance answer to recovery options and avoid interpreting QuotaLimit as token-only exhaustion.
• Durable guidance: update AGENTS.md and .agents/references/terminology.md with the reader-value test, canonical pricing ownership, natural plan names, connected prose, exact shipped labels, scoped credential claims, source-versus-rollout evidence, and cross-page propagation of review comments.
• Final platform-billing follow-up: remove a universal setup-time statement so harness-specific rules define when billable activity starts.
• Final Enterprise-billing follow-up: state that Enterprise spending limits cap usage charges, including pay-as-you-go usage, rather than purchase transactions.
• Final handoff follow-up: remove the unsupported model-preservation and Auto fallback statement while retaining the verified credential boundary.
• Final pricing-source follow-up: remove the unsupported 20% Free-plan purchase premium from three pages while retaining qualitative premium guidance and current-price links.

Content design plan

Reader and job: An existing customer checking an allowance or converted balance, buying usage, or investigating run costs.
Gap today: Credit-era explanations conflate usage, payment amounts, recorded charges, and estimates.
Change: Update existing billing owners and linked summaries; exclude internal mechanisms, future proposals, and unconfirmed cohort promises.

Unverified claims

• Live rollout and balance/contract migration — checked-in production configuration now specifies activation, superseding the earlier unset observation. Source alone does not confirm deployed availability, stable-client support, or migration of existing balances or contracts. Usage guidance relies on the billing unit already shown to the reader and does not assert availability of a display control. The conditional self-serve conversion rule is source-backed below; no examples or universal migration are promised.
• Allowance cohorts and grandfathering — the proposed blanket Max and legacy Business amounts conflict with source configuration. The disputed table and permanent-retention promises are omitted; confirm approved amounts, eligibility, and plan-change effects before adding them.
• Enterprise POC migration and detailed reporting availability — no universal POC conversion, detailed API reporting, or web-view availability is promised. Confirm contract scope and the live reporting surfaces; conditional UI guidance remains on the usage and factory pages.
• Remaining purchase-policy discrepancy — confirm eligible legacy tiers and the customer-facing auto-reload/purchase-limit policy. Purchase sizes and reset-cycle distinctions are source-backed; docs do not describe the limit as a guaranteed ceiling on charged amounts.
• Concurrency policy confirmation — the existing “still being finalized” placeholder remains in src/content/docs/platform/faqs.mdx. Product and Billing must confirm limits and allocation details before replacing it; the consistency seed and tests were not changed to bypass that gate.

Documentation risk

Risk: engineering-review-required
Rationale: Changes billing, allowance, purchase, reporting, and credential claims; rollout and policy confirmations remain pending. The fourteen-page follow-up removes 1,020 words and five callouts; the overall fifty-nine-page update removes 6,377 words and twenty-five callouts against the comparison baseline below. BYOK, personal endpoints, and team-managed inference now meet the 1,500-word and two-callout budgets. CLI configuration retains its existing overage at 1,711 words, including six words added in this follow-up to identify the shipped control. The pricing FAQ retains four callouts. Full Terminal Use remains at 1,956 words to preserve the requested explanatory prose; it is unchanged in this follow-up. These and other inherited setup, security, and integration findings remain review decisions, not reasons to cut unrelated procedures. No page newly crosses its applicable length budget, and callout counts do not increase.
Source files consulted: warp/app/src/settings_view/billing_and_usage_dispatch.rs@5c84259e2e4ba7c331cabac02649cff049922822, warp/app/src/settings_view/billing_and_usage_page_v2.rs@5c84259e, warp-server/logic/billing.go@a519280bd459e78ed057fcb922966d11ec2c85b8, warp-server/logic/ai/ambient_agents/billing/billing.go@a519280b, warp-server/logic/ai/ambient_agents/billing/platform.go@a519280b, warp-server/logic/ai/ai_usage/cost_resolution.go@a519280b, warp-server/config/prod.yaml@a519280b, warp-server/model/types/ai_request_bonus_grants.go@a519280b, warp-server/model/types/ai_request_bonus_grants_test.go@a519280b, warp-server/logic/ai/grants/grants.go@a519280b, warp-server/logic/ai/llm/llm.go@a519280b, warp-server/model/ai_request_bonus_grants.go@a519280b, warp-server/logic/api_keys.go@a519280b, warp-server/logic/ai/ai_usage/usage.go@a519280b, warp-server/router/handlers/public_api/agent_webhooks.go@a519280b, warp-server/logic/ai/ambient_agents/dispatcher.go@a519280b, warp-server/specs/pricing-transparency/free-adhoc-credit-purchases/PRODUCT.md (5-18, 32-43)@a519280b, warp/app/src/lib.rs@721449fde71e6541a82e7b66d8682960b4df450f, warp/app/src/settings/ai.rs@721449fd, warp/crates/warp_tui/src/usage.rs@721449fd, warp/app/src/ai/agent/conversation.rs@721449fd, warp/app/src/server/server_api.rs@721449fd, warp/app/src/ai/llms.rs@721449fd, warp-server/logic/ai/team_byo_inject.go@fb3aed1890f775c8859e8fce3f088a734864442d, warp-server/router/handlers/generate_multi_agent_output.go@fb3aed18, warp-server/logic/organization_settings.go@fb3aed18, warp-server/logic/ai/ai_usage/spending_limits.go@fb3aed18, warp-server/logic/ai/llm/llm_output/util.go@fb3aed18
Requested engineering reviewers: none
Engineering review status: pending
Docs override: none

Validation

Fresh local checks below ran on the integrated source published at 83ee1257, superseding the earlier heads' missing-dependency and pending-CI notes.

• Node 22.20.0: the required ordered typecheck, five heading-variable tests, build, rendered TOCs, and homepage JSON-LD all passed. Typecheck reports 0 errors, 0 warnings, and 7 existing hints; TOC validation checked 3,356 entries across 396 HTML files.
• All 20 current CI Python/self-test commands passed, including the audit, quality-policy, publication, UI-reference, body-integrity, reviewer-routing, and redirect suites.
• UI references: 59 pages, 0 issues, using the refreshed snapshot (warpdotdev/warp@fc4b06fb46ce8938923748a1a0ade7995a429efa). Internal links: 4,364 checked, 0 broken. Built-HTML verification preserves 422 pre-review anchors across 26 pages, including both monthly-limit anchors and Full Terminal Use's #usage and #credits-usage. The explanatory prose and exact concurrency placeholder remain intact. Diff-whitespace checks passed.
• Style audit: 271 advisories, 0 errors across 59 pages: 220 unrecognized-term suggestions, 23 screenshot-width warnings, 8 meta-openers, 7 buzzwords, 6 heading-case warnings, 4 consecutive-callout warnings, and 3 callout-density warnings.
• Compression: the fourteen-page follow-up removes 1,020 words and five callouts. Overall, 6,377 words and twenty-five callouts are removed across 59 pages. BYOK is 1,212 words, personal endpoints 1,323, and team-managed inference 1,478; each has two callouts. CLI configuration's existing word overage, the pricing FAQ's four callouts, Full Terminal Use's requested explanatory prose, and other inherited findings are acknowledged above. No new budget crossings or callout growth.
• Informational production dependency audit: five existing vulnerabilities (three high, two low). Dependencies are unchanged; no unrelated audit fixes were attempted.
• Trunk is unavailable. External-link checks and GUI verification were not performed; no current approved replacement billing screenshots were available.
• At 623e27e3, 4,364 internal links passed with none broken, changed UI references had no issues, style lint exited 0 with 271 advisories, and the additional-usage and agent FAQ pages passed compression. Diff whitespace passed. All exact-head CI passed, including Agent docs review with 0 critical, 0 important, 4 suggestions, and 0 nits. All thirteen existing review threads remain resolved, including the ten human threads with eleven comments. Earlier independent reviews removed or conditioned their important claim-verification findings. Read-only source audits found no remaining blockers, but source tests were inspected, not executed. Engineering, deployment, and policy confirmation remain pending.

Review routing and provenance

Ownership resolution returned multiple human candidates (pkillianjackson, tylerlam-warp, abhishekp106, and jefflloyd), so no reviewer or team is requested. Engineering and release/policy review remain pending.

• Conversion rule — plans-pricing-refunds.mdx, “Existing purchased balances”: at the pinned server revision, model/types/ai_request_bonus_grants.go (66-76) supports the conditional higher-of rule. The default rate is in logic/ai/llm/llm.go (2853-2862), the valuation call is in logic/ai/grants/grants.go (192-203), and proportional remaining balances are in model/ai_request_bonus_grants.go (384-403). Public conversion examples were removed because migration availability remains unconfirmed.
• Free-plan purchased usage — specs/pricing-transparency/free-adhoc-credit-purchases/PRODUCT.md (5-18, 32-43) supports purchases without a subscription, 12-month expiry, and team-scoped access. The team pages retain the existing paid-seat proration rule and link to its billing owner.
• Agent API key billing — platform/team-access-billing-and-identity.mdx and its billing/API-key/error summaries: warp-server/logic/api_keys.go (611-641) and router/handlers/public_api/agent_webhooks.go (899-913) establish the authenticated agent; logic/ai/ambient_agents/billing/billing.go (21-59) and logic/ai/ai_usage/cost_resolution.go (231-247) select the billed account. logic/ai/ai_usage/usage.go (142-149, 379-408) and model/ai_request_bonus_grants.go (214-265) support eligible shared usage rather than the owner's personal included allowance. Schedule defaults are separately supported by logic/ai/ambient_agents/dispatcher.go (2675-2691). All references use the pinned server revision above.
• Cross-page inference audit — at client 721449fde71e6541a82e7b66d8682960b4df450f and server fb3aed1890f775c8859e8fce3f088a734864442d, warp/app/src/lib.rs (532-554) distinguishes app and CLI credentials; warp-server/logic/ai/team_byo_inject.go (46-137) and router/handlers/generate_multi_agent_output.go (161-217, 2247-2350) support team-managed selection and injection; warp/app/src/ai/llms.rs (1335-1382) supports cloud-model fallback. Storage, transit, provider retention, Auto, and billing claims remain scoped to their configuration.
• Quota and reported-cost audit — at the same revisions, the handler cited above (1425-1426) and logic/ai/llm/llm_output/util.go (49-54) show that QuotaLimit is not exclusively token exhaustion. warp/crates/warp_tui/src/usage.rs (82-90) and app/src/ai/agent/conversation.rs (220-257) distinguish reported Warp charges from external provider cost. Related source tests were inspected, not executed; no live customer or deployment verification is claimed.

Source revisions are pinned above. Docs comparison baseline: 9be490643fe996fb6ed8ab119588a675d5b8aa28; synced main: ce21ea93; current docs head: 623e27e39533c9fc46811e7dd9a7d2d77c740193. This follow-up preserves the human review revision at 83ee1257dc5a6b3f49f13d5bae9696aa8f05e18f and the platform-billing correction at 3830bc9df3d9a65178e415aa9a679d62048e8e66. Existing overlapping PRs #833 and #839 were checked and left unchanged. The PR’s ready-for-review state, reviewer list, labels, and existing attribution comment remain unchanged; engineering and policy confirmation remain pending.

Co-Authored-By: Warp agent@warp.dev

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated
docs	Ready	Preview	Oct 10, 2026 2:25am UTC

[hongyi-chen]

This PR was generated with Warp.

Comment @warp-agent on this PR to send it follow-up work.

 

[warp-local-for-testing-only]

This repository is not enabled for a Warp team. Ask a Warp workspace admin to configure GitHub repository access in the Warp admin panel.

[warp-for-oss]

@hongyi-chen

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 /warp-agent-review on this pull request to retrigger a review (up to 3 times on the same pull request).

Powered by Oz

[warp-for-oss]

Overview

This PR updates existing billing, usage, inference, cloud agent, factory, and Enterprise docs to distinguish dollar-billed usage from credit-based agreements while preserving existing URLs and legacy anchors.

Concerns

No blocking issues found in the attached diff. The provided spec context contains no approved or repository spec context to compare against, and the security pass found no changed code or configuration with security impact.

Verification

• Internal link check passed: 4,358 internal links checked, 0 broken.
• Documentation risk contract check passed: 0 VERIFY markers across 58 changed public docs files.

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

[github-actions]

Review summary

The independent agent completed its review for this commit.

Findings

• src/content/docs/support-and-community/plans-and-billing/platform-credits.mdx:20 — [IMPORTANT] The new sentence hard-codes "The standard platform rate on dollar-billed self-serve plans is $5 per agent hour" and deletes the base version's "See Warp pricing for current rates." pointer; a literal rate drifts silently and the claim is absent from the PR's ## Unverified claims. Requested change: restore the pricing-page deferral, or keep the rate and list it under ## Unverified claims for explicit engineering confirmation.
• src/content/docs/enterprise/support-and-resources/billing.mdx:36 — [IMPORTANT] The rewritten "How credits are consumed" paragraph leaves line 38 stale: "For agent API key runs (e.g., CI/CD pipelines, scheduled tasks), credits also draw from the team's shared pool since these runs are not tied to any individual user." This contradicts the new canonical rule in platform/team-access-billing-and-identity.mdx that user-created schedules bill the creator by default. Requested change: separate agent API key runs from user-created schedules on line 38, matching #billing-for-cloud-agent-runs, or drop the example list and link to it.
• src/content/docs/support-and-community/plans-and-billing/credits.mdx:17 — [IMPORTANT] "On dollar-billed plans" is the first of 24 uses of a qualifier this PR introduces across the docs set, but no page tells readers how to determine whether their plan is dollar-billed, and "Dollar-billed Free plans" implies a non-dollar-billed Free plan exists. Requested change: define the dollar-vs-credit distinction once in this section, say where the reader checks which applies to them, and drop the repeated qualifier from downstream sentences the definition already covers.
• src/content/docs/support-and-community/plans-and-billing/platform-credits.mdx:22 — [SUGGESTION] The replacement keeps the queued/idle exclusions but drops the base version's accrual window: "A run accrues agent hours from the time the agent starts in its environment, including environment setup, until the run finishes, fails, or is cancelled." Requested change: restore the start/stop boundary, including that environment setup counts and that failed and cancelled runs still accrue.
• src/content/docs/support-and-community/plans-and-billing/add-on-credits.mdx:59 — [SUGGESTION] The paragraph states the limit governs "One-time purchases and auto-reloads," but it sits under #### Configuring a monthly spend limit nested inside ### 2. Enable auto-reload, hiding it from on-demand buyers; the heading also still says "spend limit" while the body and the new glossary entry prefer "monthly purchase limit." Requested change: promote it to a sibling ## section after both purchase paths and preserve the inbound #configuring-a-monthly-spend-limit anchor with a <span id> shim.
• src/content/docs/platform/index.mdx:133 — [SUGGESTION] "You need enough available usage to start the run" replaces the checkable "at least 20 credits available" in a prerequisites list, and the same vague substitution lands in platform/index.mdx:145, platform/quickstart.mdx:34, platform/integrations/jira.mdx:18, slack.mdx, linear.mdx, and scheduled-agents-quickstart.mdx. Requested change: after engineering confirms the current threshold, state the minimum or say plainly there is no fixed minimum, and point the reader to Settings > Billing and usage.
• src/content/docs/agents/capabilities/full-terminal-use.mdx:195 — [SUGGESTION] The rewrite dropped the "To manage credit usage:" lead-in, so the three mitigation bullets on lines 197-199 dangle off a sentence about why interactive tasks cost more. Requested change: end the paragraph with a lead-in such as "To keep usage down:" before the list.
• src/content/docs/guides/configuration/how-to-use-tokens-efficiently-with-ai-coding-agents.mdx:13 — [SUGGESTION] This Guides tutorial lost both ends of its required structure: the opening no longer states what the reader will accomplish, and ## Next steps at line 87 is now a bare link list with the conclusion paragraph deleted, against AGENTS.md's "End with a conclusion, then next steps." Requested change: restore a one-sentence goal statement here and a one-to-two-sentence recap above the Next steps links.
• src/content/docs/terminal/settings/all-settings.mdx:269 — [SUGGESTION] The PR body's risk rationale claims "No new budget violation, worsened word overage, or increased callout count was introduced," but this file grew 4,906→4,918 words, agents/cli/configuration.mdx 1,709→1,719, and platform/index.mdx 1,506→1,507, all already over the 1,500-word budget. Requested change: amend the rationale to note that three already-over-budget pages grew by 1-12 words for cross-references.
• src/content/docs/support-and-community/plans-and-billing/pricing-faqs.mdx:175 — [SUGGESTION] The answer is fully converted to usage/purchase language while its H3 on line 173 still reads "Who on my team can purchase add-on credits?" (and line 179 likewise), even though this PR renames headings behind <span id> shims everywhere else. Requested change: either apply the same shim-and-rename treatment, or state in the PR body that FAQ headings intentionally keep user vocabulary.
• src/content/docs/agents/inference/bring-your-own-api-key.mdx:98 — [SUGGESTION] Neither this bullet nor the one below it mentions credits, but the heading on line 94 is still ### Credit usage, unlike custom-inference-endpoint.mdx where ### AI credits became ### Inference usage behind a <span id="ai-credits" /> shim. Requested change: add <span id="credit-usage" /> and rename the heading to ### Inference usage so it matches its body and keeps the anchor resolving.
• src/content/docs/agents/inference/grok-subscription.mdx:91 — [NIT] The new "Usage and billing" list item uses a hyphen separator while the four items above it use an em dash; the same mismatch appears at bring-your-own-api-key.mdx:157, custom-inference-endpoint.mdx:145, and insufficient-credits.mdx:60-63. Requested change: use an em dash on the new item to match each list's existing style.

Verdict

Request changes

[warp-agent-staging]

Overview

This PR substantially expands pricing and usage documentation. Two points need human judgment before the final pass: the published Free-plan purchase path and whether Product and Billing approved removing the cloud-concurrency placeholder.

Concerns

• The canonical purchase instructions include Free-plan readers but point them to a desktop control that the cited product spec says is unavailable to them; choose plan-scoped instructions or the Free web purchase path.
• The concurrency-placeholder removal triggers an intentional owner-confirmation audit; record the approval and resolve the audit seed/tests, or retain the placeholder.

Verdict

Checks: build fail, tests fail, CI red, visual proof n/a

Found: 0 critical, 2 important, 0 suggestions, 0 nits

Responding as Docs Factory (V2): Open session · View in factory

[warp-local-for-testing-only]

@hongyi-chen, your GitHub account is not connected to Warp. Connect your GitHub account.

[warp-agent-staging]

The revision resolves both earlier pricing-transparency findings. One broken in-page link remains in the auto-reload procedure, so this revision still needs a small correction.

Found: 0 critical, 1 important, 0 suggestions, 0 nits

Responding as Docs Factory (V2): Open session · View in factory

[warp-agent-staging]

The current revision resolves the broken anchor and preserves the earlier pricing-surface and concurrency-placeholder corrections. I found no remaining issue in the PR-owned diff; required build and agent-review checks are still pending.

Responding as Docs Factory (V2): Open session · View in factory

[github-actions]

Review summary

The independent agent completed its review for this commit.

Findings

• src/content/docs/support-and-community/plans-and-billing/plans-pricing-refunds.mdx:26 — [IMPORTANT] The credit-pack conversion rule and the $10/$117 examples (and the "Most plans show allowances and balances in dollars" sentence at line 22) are stated as fact, while the PR body's Unverified claims says live rollout and balance migration are unconfirmed and no VERIFY marker exists. Requested change: get Billing/Engineering to confirm rollout and migration before merge, or scope the wording conditionally, or add VERIFY markers and list them under ## Unverified claims.
• src/content/docs/platform/team-access-billing-and-identity.mdx:39 — [IMPORTANT] "Eligible pay-as-you-go agreements allow runs without a prepaid balance" is a new Enterprise billing claim with no traceable cited source, and it does not say which agreements qualify. Requested change: cite the supporting source and name the qualifying condition, or remove the sentence and point readers to their contract or Enterprise billing.
• src/content/docs/support-and-community/plans-and-billing/credits.mdx:105 — [SUGGESTION] The sentence "Runs started with an agent API key use eligible shared team usage, not the team owner's personal included allowance. Cloud agents don't receive a separate per-seat allowance." is repeated nearly verbatim in platform/team-access-billing-and-identity.mdx:221 and agents/cli/oz-cli/api-keys.mdx:86. Requested change: keep the full statement on one canonical page and replace the other two with a one-line summary and link.

Verdict

Request changes

[github-actions]

Review summary

The independent agent completed its review for this commit.

Findings

• src/content/docs/knowledge-and-collaboration/teams.mdx:74 — [IMPORTANT] The edit reverses the published rule that purchased usage requires an active subscription (downgrading to Free suspended access) and drops the sentence about the admin paying a prorated seat rate on rejoin, with no cited source and no entry under ## Unverified claims. Requested change: cite the source file and lines supporting the new behavior (or list it as unverified), and restore the seat-proration sentence or confirm it is obsolete; apply the same fix to src/content/docs/enterprise/team-management/teams.mdx:76.
• src/content/docs/support-and-community/plans-and-billing/credits.mdx:54 — [SUGGESTION] "Dedicated cloud allowances can be used before your general balance" is never defined, and the same phrase appears in add-on-credits.mdx:66, pricing-faqs.mdx:141, platform/team-access-billing-and-identity.mdx:224, and insufficient-credits.mdx:11. Requested change: define the term once (what it is, who gets it, where it shows in Billing and usage) and link to it, or remove the caveat where it isn't actionable.
• src/content/docs/support-and-community/plans-and-billing/credits.mdx:22 — [SUGGESTION] The clause "rather than inferring the billing unit from a displayed amount" argues with an imagined reader; the dropdown claim itself is supported by the public client. Requested change: delete the clause and keep the instruction to check billing terms or contact billing@warp.dev.
• src/content/docs/support-and-community/plans-and-billing/credits.mdx:30 — [SUGGESTION] "Warp-provided inference is priced at public API rates" (also at src/content/docs/agents/getting-started/faqs.mdx:46) is a pricing claim cited only to a whole file and not listed under ## Unverified claims. Requested change: add the specific file and line range to Source files consulted, or link to the Warp pricing page instead of restating it.
• src/content/docs/support-and-community/plans-and-billing/add-on-credits.mdx:32 — [SUGGESTION] The line defers to Warp pricing for amounts, then hard-codes "Free-plan purchases include a 20% premium," repeated at pricing-faqs.mdx:165 and agents/getting-started/faqs.mdx:48. Requested change: state the premium once on the owning page (or only on the pricing page) and have the FAQ pages link to it.
• src/content/docs/agents/inference/bring-your-own-api-key.mdx:156 — [SUGGESTION] Link-only Related pages items were newly wrapped in bold (**[Custom inference endpoint](...)**), which adds no emphasis and differs from sibling pages. Requested change: remove the bold wrappers here and in custom-inference-endpoint.mdx, grok-subscription.mdx, and insufficient-credits.mdx.
• src/content/docs/agents/capabilities/full-terminal-use.mdx:193 — [SUGGESTION] The page grows from 1,936 to about 1,956 words against the 1,500-word feature-doc budget; the PR body justifies the overage. Requested change: run a deletion-only pass on the unrelated procedures at the next edit so the overage shrinks.

Verdict

Request changes

[github-actions]

Review summary

The independent agent completed its review for this commit.

Findings

• src/content/docs/agents/getting-started/faqs.mdx:46 — [IMPORTANT] "Warp-provided inference is priced at public API rates" is stated for all self-serve paid plans, but credits.mdx scopes it to plans billed in dollars and the PR body says the follow-up scoped it. Requested change: scope this sentence to dollar-billed self-serve plans or link to /support-and-community/plans-and-billing/credits/#dollar-usage-and-credits instead of restating it.
• src/content/docs/agents/getting-started/faqs.mdx:69 — [IMPORTANT] The QuotaLimit explanation changed from an AI token limit to "A usage quota has been reached" with no cited source and no entry under Unverified claims. Requested change: cite the server source that returns QuotaLimit, or restore the original wording and list the change as unverified.
• src/content/docs/agents/cli/models-and-usage.mdx:48 — [IMPORTANT] Claims that /cost does not follow the statusline's dollar-display preference, and that the statusline switches units when both are available, describe CLI behavior, but the cited sources are desktop billing files and server code. The same applies to usage_display_mode at src/content/docs/terminal/settings/all-settings.mdx:269 and src/content/docs/agents/cli/configuration.mdx:83. Requested change: add the CLI source files to Source files consulted, or revert to the previous wording and list the claim under Unverified claims.
• src/content/docs/factories/factory-dashboard.mdx:48 — [IMPORTANT] New claims that a run's Usage under Metadata shows per-model and per-component cost, that incomplete totals are not shown, and that the Warp app display preference does not change web units have no cited web client source, and the PR body says the live reporting surfaces are unconfirmed. Requested change: cite the web client source for the Usage panel or reduce the text to what the cited server code supports.
• src/content/docs/enterprise/support-and-resources/billing.mdx:16 — [SUGGESTION] Headings "How credits work" (line 16), "Credit allocation" (line 22), and "How credits are consumed" (line 34) still use credit terms although the page now frames usage as contract-specific. Requested change: rename them to usage terms and keep the old anchors with span ids as elsewhere in the PR.
• src/content/docs/support-and-community/plans-and-billing/pricing-faqs.mdx:209 — [SUGGESTION] The "review purchase controls before relying on unattended runs" caution repeats in this file, agents/cli/oz-cli/api-keys.mdx:89, platform/index.mdx:142, and platform/team-access-billing-and-identity.mdx:237. Requested change: keep it once in platform/team-access-billing-and-identity.mdx#billing-for-cloud-agent-runs and link to it from the others.
• src/content/docs/support-and-community/plans-and-billing/plans-pricing-refunds.mdx:42 — [SUGGESTION] The refund condition changed from "no credits have been used" to "no usage has been consumed" (also line 48 and in pricing-faqs.mdx), which widens the policy wording while the PR body says refund policies were preserved. Requested change: confirm the wording with Billing or restore the original text.
• src/content/docs/guides/configuration/how-to-use-tokens-efficiently-with-ai-coding-agents.mdx:89 — [SUGGESTION] "Compare your next conversation with your usage baseline..." is a leftover recap line at the end of the Planning section. Requested change: delete it or move it into an actionable step under Track your usage first.
• src/content/docs/agents/inference/bring-your-own-api-key.mdx:64 — [SUGGESTION] The PR removes references to Pricing-Blog-BYOK.png, reload-credits.png, inline-credit-usage-footer.png, and usage-summary-panel.png, but the four files remain in src/assets/ unreferenced. Requested change: delete the unused PNGs in this PR.
• src/content/docs/factories/infrastructure-and-security.mdx:144 — [NIT] The Platform usage bullet lost its bold lead term while sibling bullets keep theirs. Requested change: use * [**Platform usage**](/support-and-community/plans-and-billing/platform-credits/) - Understand agent-hour charges.

Verdict

Request changes

[github-actions]

Review summary

The independent agent completed its review for this commit.

Findings

• src/content/docs/enterprise/support-and-resources/billing.mdx:71 — [IMPORTANT] The sentence "Spending limits apply to both purchased usage and pay-as-you-go usage." contradicts the note at line 74, "These Enterprise limits cap usage, not purchases." Readers cannot tell whether the Enterprise cloud, local, total, and per-user limits cover purchases. Requested change: confirm the behavior against spending_limits.go, then reword both lines so they state one consistent rule, for example that the limits cap consumption of usage however it is funded and are separate from the self-serve monthly spend limit on purchases.
• src/content/docs/enterprise/support-and-resources/billing.mdx:34 — [SUGGESTION] The headings "How credits work" (line 16), "Credit allocation" (line 22), and "How credits are consumed" (line 34) still use credit framing, but their content now describes usage in the contract's billing unit. This conflicts with the page's own new text that credits are only one possible unit. Requested change: rename them to usage-based headings such as "How inference usage is measured", "Usage pools", and "How usage is consumed", and keep any inbound anchors with <span id> shims as done elsewhere in the PR.
• src/content/docs/agents/capabilities/slash-commands.mdx:24 — [SUGGESTION] The /cost row in the table (line 21, "Toggle credit usage details in the current conversation.") still says credit usage, while models-and-usage.mdx:50 now describes /cost as hiding or showing elapsed time and usage. The same page's changed caution at line 24 already moved to "metered inference". Requested change: reword the /cost row to match, for example "Show or hide usage details for the latest response."
• src/content/docs/support-and-community/plans-and-billing/pricing-faqs.mdx:165 — [SUGGESTION] The "20% premium" for Free-plan purchases is hardcoded in three places: this line, add-on-credits.mdx:32, and agents/getting-started/faqs.mdx:48. The new AGENTS.md guidance says current prices belong on the pricing page and FAQs should summarize and link to the owning page. A repeated rate will drift. Requested change: keep the figure only on add-on-credits.mdx, and have the two FAQ entries say purchases carry a Free-plan premium and link to the purchase page.
• src/content/docs/agents/inference/bring-your-own-api-key.mdx:42 — [SUGGESTION] The PR removes the last references to src/assets/support-and-community/Pricing-Blog-BYOK.png, reload-credits.png, and inline-credit-usage-footer.png, and to src/assets/guides/usage-summary-panel.png. The files are still in the repo and nothing references them. Requested change: delete the four unreferenced image files, since the PR body says the billing screenshots are removed. Also add the reversal of agent API key billing to the PR's ## Unverified claims. Several pages now say agent API key runs use shared team usage, not the team owner's allowance (platform/team-access-billing-and-identity.mdx:221, pricing-faqs.mdx:207, insufficient-credits.mdx:11), which contradicts the prior documented behavior. Record the owner confirmation there.

Verdict

Request changes

[github-actions]

Review summary

The independent agent completed its review for this commit.

Findings

• src/content/docs/terminal/settings/all-settings.mdx:269 — [SUGGESTION] The usage_display_mode entry lists "cost" as a plain option, but the cited warp_tui/usage.rs ignores it for accounts charged in credits (they always see credits) and offers the click-to-switch only on cent-charged tiers. The PR body also says the unit-switching was removed as 'uncited', but that source supports it. Requested change: state that cost applies only to accounts billed in dollars and that credit-billed accounts always show credits. Restore a one-clause mention that the entry is clickable on dollar-billed accounts, or correct the PR body.
• src/content/docs/agents/cli/configuration.mdx:83 — [SUGGESTION] The 'Conversation usage' bullet says the chip is listed as Credit usage in /statusline, but does not say that dollar-billed accounts can see dollar totals or that the entry switches units there. Requested change: add a short clause matching the usage.rs behavior, or link to the section in models-and-usage.mdx that states it, so readers on dollar billing are not misled.
• src/content/docs/support-and-community/plans-and-billing/plans-pricing-refunds.mdx:26 — [SUGGESTION] 'If Warp converts an existing self-serve purchased credit pack to dollar usage, it keeps its face value or equivalent inference value, whichever is higher' publishes a conversion rule while the PR body says migration availability is unconfirmed. A reader cannot act on an 'if' with no stated trigger. Requested change: remove the paragraph until Billing confirms conversion is live, or state the actual trigger. Keep the instruction to check the balance shown in Billing and usage.
• src/content/docs/platform/faqs.mdx:115 — [SUGGESTION] 'Concurrency limits and credit allocation details are still being finalized.' is an unfinished-work statement in public docs, and it uses 'credit allocation' while the surrounding rewrite moves to 'usage'. The PR body lists it as a known open item. Requested change: remove the sentence or replace it with confirmed limits from Product and Billing. If it must stay, change 'credit allocation' to 'usage allocation' and link to the canonical billing page.

Verdict

Approve

[warp-agent-staging]

The final revision addresses all human review comments and removes or conditions the remaining unsupported claims. All review threads are resolved, the exact-head checks are green, and I found no remaining blocking issue in the PR-owned diff.

Responding as Docs Factory (V2): Open session · View in factory