​

Pricing model docs update — design

Date: 2026-08-01 Branch: docs/pricing-model-update, branched from mintlify-port (the production branch per README.md). PR targets mintlify-port, not main. Source of truth: the pricing model handoff brief (marketing site shipped; docs lagging).

​

Goal

Bring the Mintlify docs in line with the new commercial model: executions replace test minutes, a four-tier Free/Solo/Team/Enterprise ladder replaces the five-tier Free-Trial/Starter/Growth/Pro/Enterprise ladder, and overage billing disappears entirely.

​

Scope

One page rewritten, four leaked claims corrected, one glossary entry added. docs.json needs no change — the nav entry administration/how-quash-pricing-works and its label are still correct.

​

Explicitly out of scope

Verified as unrelated to Quash pricing; must not be swept:

• The Apple Developer Program $99/year in devices/physical-ios-devices.mdx:6 and getting-started/faqs.mdx:78.
• The Figma “free or Starter plans” reference in test-studio/recipe/attaching-context.mdx:64 — that is Figma’s tier ladder, not Quash’s.
• Credit-card payment examples in test-studio/prompting-guide-test-generation/common-mistakes.mdx — test fixtures, unrelated to billing.
• test-studio/token-management.mdx, getting-started/core-concepts/tokens-and-compact.mdx, and the glossary “Tokens” entry — these describe the agent’s context-window budget, not billing. Confirmed by grep: no metering, charging, or quota language.
• All security and compliance content (encryption, residency, retention).
• All product capability docs (what the agent does, integrations, reporting, test paths).

​

Part 1 — administration/how-quash-pricing-works.mdx

Full rewrite. Section order:

1. Frontmatter — description currently reads “Plans, test minutes, cloud sessions, overage, and what changes when you upgrade.” Replace with a description built on plans, executions, and what counts.
2. Opening — lead on the site’s framing: “Write all the tests you want. Pay only when they run.”
3. Plans at a glance — the four-tier table (below).
4. What counts as an execution — its own prominent section, ahead of tier detail. This is the definition everything else depends on.
5. No overage billing — replaces the old ## Overage section outright.
6. Free plan — replaces ## Free trial.
7. What each plan includes — “adds over the previous tier” prose.
8. Feature comparison — seven tables, recolumned (below).
9. Common questions — rewritten around executions.
10. Closing link to https://quashbugs.com/pricing (retained).

​

The hard rule

Solo is the only published price. Team and Enterprise must never show a number anywhere on the page — not a range, not a “starting at”, not an implied per-seat figure. Describe them by size:

• Team — “for teams past ~150 executions/month, 3+ seats, or multiple apps”
• Enterprise — “private deployment, 1,500+ executions/month”

Where a cost reference is unavoidable, say “quoted” or link to /pricing.

​

Plans at a glance

The brief’s separate “Card required” and “Expiry” rows are folded into the Free plan prose as “No card, no expiry” rather than carried as table rows — as rows they are only meaningful for the Free column.

​

What counts as an execution

State all three rules explicitly. This is the single most important definition on the page. One execution = one test run on a device.

• Counts wherever it runs — local device, emulator, or Quash-hosted cloud. This reverses the old “local execution doesn’t count” rule; the page must be unambiguous, because the old rule is what existing users believe. The reason to give: the cost is the agent reasoning about the app, not the device, and that is incurred identically on a local run.
• A test that fails counts the same as one that passes. Finding a failure is the product working.
• Generation never counts. Describing flows, generating test cases, refining prompts, editing suites, reading reports, inviting teammates — free on all plans including Free.

There is a fair-use rate limit on generation. It must not be documented as a number or presented as a quota.

​

No overage billing

Replaces ## Overage entirely. No plan bills overage. Use the site framing: “No overage billing. If you outgrow Solo, we’ll size a plan with you — you’ll never get a surprise invoice.” Deleted from the old page: $0.20/min, “Annual customers get 20% off overage rates”, “Overage available” as a Starter bullet, and the two overage mentions in the free-trial section.

​

Free plan

Replaces ## Free trial. 40 executions to start, then 5/month. No credit card. No expiry — the 14-day trial is gone. Full platform access minus the tier-gated capabilities below.

​

What each plan includes

• Free — unlimited test generation, local devices and emulators, prompt-to-execution testing, detailed execution reports, community support.
• Solo adds over Free: cloud devices, backend and API validation, full test suite management.
• Team adds over Solo: CI/CD integrations (GitHub Actions, GitLab, Bitbucket, Jenkins), issue-tracker integrations (Jira, Linear, GitHub Issues), webhooks and triggers, scheduled regression runs, hands-on onboarding, email support.
• Enterprise adds over Team: private VPC or on-prem deployment, custom data residency, SSO (SAML/OIDC), RBAC, SCIM provisioning, compliance documentation (SOC 2, ISO 27001, GDPR, DPDPA), dedicated real device fleet, dedicated CSM, custom SLA.

Team must not read as enterprise-only. Include the site’s framing: “Team is quoted to your actual setup — it starts wherever you are, and plenty of Team customers are small.” This framing exists to stop mid-market buyers self-selecting out; it is not optional decoration.

​

Feature comparison — row allocation

All seven tables retained, recolumned from Starter/Growth/Pro/Enterprise to Free/Solo/Team/Enterprise. Note that Free is now a column, which it was not before (the old tables started at Starter). Test Execution Test Generation Test Management Backend and API Reporting and Analytics — all three rows (detailed execution reports, AI-powered debugging insights, real-time usage tracking) stay ✓ across all tiers, unchanged from the old page. The Free column is new; all three apply to it. Integrations and Workflow Security and Administration Support Rows deleted: “Priority response (<4 biz hours)” and “Full automation guarantee”. Both are SLA-shaped promises the new model does not make; Enterprise’s “custom SLA” covers the ground.

​

Decisions on record

RBAC and SSO both land at Enterprise only. The brief places both at Enterprise but flags SSO as unverified (“an earlier internal design had SSO at a lower tier”). The current docs ship RBAC at the mid tier. User decision on 2026-08-01: follow the brief. Consequence to be aware of — an existing mid-tier customer reading the docs sees RBAC move out of reach. Six rows are judgment calls, not brief-derived facts. Marked above. Rationale:

• Adaptive execution → all tiers. Gating a core agent behaviour contradicts the model’s thesis that you are metered only on executions.
• Codebase ingestion, Figma ingestion → all tiers. The old Enterprise-only rows already contradict the rest of the docs: apps/apps-manager/knowledge.mdx gates codebase connection on nothing, and test-studio/recipe/attaching-context.mdx:64 gates Figma on a paid Figma plan rather than a Quash tier. Caveat worth restating: “generation is unlimited” is a claim about metering, and extending it to input sources is an inference. These are the two rows most worth a second opinion before publish.
• Database validation → Team+. The brief names Solo’s adds as “backend and API validation” — pointedly not database — and the old page gated it above the $99 tier.
• Agentic app monitoring → Team+. Absent from the brief; Team is now the second-highest tier as Pro was.

​

Common questions

Rewrite the section around executions. Questions to carry:

• What counts as an execution? (restate the definition)
• Do local runs count? — yes, explicitly reversing the old rule
• Does a failed test count? — yes, same as a pass
• Does generating tests count? — no
• What happens when I run out of executions? — no overage; size up
• When should I move from Solo to Team? — the size-based framing, no number
• Can I add more apps?
• Cloud devices vs the real device fleet? — fleet is Enterprise, no longer a Pro add-on
• How does annual billing work? — Solo $99 →$79/mo. Delete the old Growth/Pro annual figures and the “20% off overage” clause.
• How does the Free plan work? — 40 then 5/month, no card, no expiry
• Can I bring my own LLM API keys? — Enterprise (retained)
• What is in Enterprise that I cannot get in Team? (was “…in Pro”)
• I am a fintech or regulated company — which plan? — Enterprise (retained, update tier names)
• I am already on a custom plan — what happens? — retained verbatim, still true and more relevant than ever during a pricing migration

​

Strings that must not survive

Verify by grep after the rewrite: test minute, agent minute, cloud device minute, Starter, Growth, $399, $899, $319, $719, $0.20, overage (except in the “no overage billing” construction), 14-day, 14 days, 100 test minutes, concurrent cloud session, free credits, credit balance, recharge, add-on on Pro. Two of these have legitimate survivors elsewhere in the repo, so scope the grep accordingly rather than sweeping blindly:

• Starter must still appear in test-studio/recipe/attaching-context.mdx:64 — that is a Figma tier. Grep for it only within the five in-scope files.
• $99 must still appear in devices/physical-ios-devices.mdx:6 and getting-started/faqs.mdx:78 — Apple Developer Program — alongside its legitimate new Solo uses.

Pro as a bare word also survives legitimately in unrelated contexts (MacBook Pro, Gemini 2.5 Pro, Apple Developer Program); only tier-name uses are in scope.

​

Part 2 — leaked plan claims

Per user decision: soften, don’t delete. Drop the numbers and per-tier framing, keep a true statement that cloud device capacity depends on plan. Users still need an answer for why they hit a queue.

​

Part 3 — glossary

Add an Execution entry to reference/glossary.mdx, positioned between “Environment tag” (line 39) and “Guidance” (line 43) to hold the file’s rough alphabetical order. Follows the file’s existing shape: ## Heading, definition paragraph, then a [→ link] line pointing to /administration/how-quash-pricing-works. Content: one test run on a device; counts on local, emulator, and cloud alike; failures count the same as passes; generation does not count. No “test minute” entry exists in the glossary, so there is nothing to remove.

​

Amendment — MCP availability (2026-08-01)

Added after the initial pass, on user report that MCP had been left out of the plan model entirely. The MCP server is available on every plan, including Free. It is not a paid upgrade. Plan limits still apply to what you do through it: agent-started tests count as executions like any other run, and plan-gated capabilities (cloud devices, for instance) stay gated when the request arrives over MCP. Reflected in six places:

• Plans at a glance — new MCP server row, ✓ across all four tiers.
• What counts as an execution — the first rule broadened from “wherever it runs” to “wherever it runs, and however you start it”, naming the Quash app, web console, and MCP as equivalent. This closes a gap the original draft left: a user could have read the execution definition as covering device type only, and assumed agent-driven runs were metered differently or not at all.
• What each plan includes — added to the Free bullet list, so it inherits up the ladder.
• Integrations and Workflow table — new row, ✓ across all tiers. It is the only row in that table not gated at Team, which is the correct and useful contrast.
• Common questions — new “Is the Quash MCP server available on my plan?” entry carrying both halves: free to connect, limits still apply.
• Glossary — the Execution entry now names how a run is started alongside where it runs.

Plus one edit outside the pricing page: quash-mcp/overview.mdx “Before you install” now says “A Quash account on any plan, including Free”, linking to the pricing page. Someone evaluating MCP reads that page, not the pricing page.

​

Deliberately not done

Per §5 of the brief, nothing on the pricing page asserts per-surface metering behaviour. The product spans three surfaces with different meters — platform already meters runs, while console and desktop were still on test minutes and migrating. The page stays surface-agnostic: it describes the published commercial model without claiming what any given surface displays in-product. Resolved 2026-08-01: the user confirmed the surface metering gap is being closed within a day, so no migration note is needed. The surface-agnostic framing stands on its own merits and needs no follow-up.

​

Open items to raise at handoff

Carried from §5 of the brief; none block this work, all are worth confirming before or shortly after publish:

1. Is test generation genuinely unmetered in console today? The old playground granted 3 free prompts per generation. The rewritten page asserts unmetered generation on every plan.
2. Is the 14-day trial expiry fully removed everywhere in-product? The page now states “no expiry”.
3. Does any surface still bill overage? The page now states no overage on any plan.
4. SSO/RBAC at Enterprise — decided above, but the brief flags SSO’s tier as genuinely uncertain.
5. The codebase/Figma ingestion rows, per the judgment-call note above.

​

Verification

• grep for every string in “Strings that must not survive”; confirm only the sanctioned survivors remain.
• grep -rn 'Team\|Enterprise' across the pricing page for any adjacent currency figure — Solo must be the only priced tier.
• Confirm the out-of-scope files are untouched: git diff --stat should list exactly the five files in Scope.
• mint dev renders the pricing page without MDX errors; all tables well-formed; the  /< escaping conventions in the existing files preserved.
• Internal links remain absolute root-relative per README.md.