Watch a run
A real run of Dummy Run against this website. Step through what the persona saw and did.
This is a real run of Dummy Run against dummyrun.dev. On 19 October 2026 the persona priya-qa-lead read this website in Firefox. The job was to work out what Dummy Run does and how a finding is checked. A model chose every action.
The persona gave a pass verdict. It also said it found browser details in the schema and example, but no dedicated list of supported browsers.
The data is a trimmed copy of one run record with the screenshots resized. The format is in the run record reference.
The run
- Persona
- priya-qa-lead (desktop)
- Scenario
- Does it replace my scripted suite
- Date
- Model
- gpt-5-4-mini-global (global)
- Browser
- firefox 158.0
- Steps
- 24
- Tokens
- 242,949 in, 1,247 out (23 model calls)
- Cost
- $0.15 Worked out afterwards from the token counts and the engine's price table, because the record was written before this model was priced.
- Outcome
- completed, verdict pass
- Run ID
2026-10-19-priya-qa-lead-does-it-replace-my-suite-31099a2c8c853e8f
What it found
The reviewer is not built yet. Every finding below is the persona's own claim and still pending.
The home page and How it works explain Dummy Run as a persona-based exploratory testing tool
The docs say the runner writes the run record and the reviewer completes it, which is the closest thing to a verification gate
The run record example shows Firefox as a supported browser in the recorded run
I found browser details in the schema/example, but not a dedicated supported-browsers list
The steps
Step 1: visit https://dummyrun.dev/

- Page
- Dummy Run https://dummyrun.dev/ · 200 · 656 ms
- Action
- visit, target https://dummyrun.dev/
Accessibility outline the persona read
link "Skip to content" @1 banner link "Dummy Run" @2 [current] navigation "Site" link "How it works" @3 link "Who it is for" @4 link "Persona library" @5 link "Docs" @6 link "GitHub" @7 link "Get started" @8 button "Switch to light theme" @9 main heading 1 "AI plays your users and tries your app before your customers do." link "Get started" @10 link "See how a run works" @11 region "Start with your job" heading 2 "Start with your job" link "QA lead Every role, every pull request" @12 link "Developer Your laptop or CI" @13 link "Agency Every client, in its market" @14 link "Product and UX Friction, with the screenshot" @15 link "Compliance and security What each role can reach" @16 link "Executive One board, with evidence" @17 region "How a run works" heading 2 "How a run works" link "Read how it works" @18 region "What you get back" heading 2 "What you get back" region "A persona library for every site" heading 2 "A persona library for every site" link "Browse the persona library" @19 region "Straight answers" heading 2 "Straight answers" link "Beside your scripted suite" @20 link "The browser table" @21 link "What goes where" @22 link "Get started" @23 region "Every run in Grafana" heading 2 "Every run in Grafana" link "See the dashboards" @24 ...Step 2: look

- Page
- Dummy Run https://dummyrun.dev/ · 200 · 285 ms
- Persona said
- I read the home page and confirmed the product positioning: Dummy Run is about running personas through scenarios against a preview, with a runner and a reviewer in the loop.
- Action
- look
- Findings
- The home page and How it works explain Dummy Run as a persona-based exploratory testing tool
Step 3: click Read how it works

- Page
- How it works | Dummy Run https://dummyrun.dev/how-it-works/ · 200 · 821 ms
- Persona said
- I followed the How it works path to understand the flow.
- Action
- click, target Read how it works
- Findings
- The home page and How it works explain Dummy Run as a persona-based exploratory testing tool
Accessibility outline the persona read
link "Skip to content" @1 banner link "Dummy Run" @2 navigation "Site" link "How it works" @3 [current] link "Who it is for" @4 link "Persona library" @5 link "Docs" @6 link "GitHub" @7 link "Get started" @8 button "Switch to light theme" @9 main navigation "Breadcrumb" link "Home" @10 heading 1 "How it works" heading 2 "Targets and models" heading 2 "Evidence" heading 2 "Words we use" link "persona library" @13 link "mascot" @14 link "How the name was chosen" @15 complementary "Next step" link "Get started" @16 link "site/content/how-it-works.md" @17 contentinfo navigation "Footer" heading 2 "Product" link "How it works" @18 link "Who it is for" @19 link "Persona library" @20 link "Get started" @21 heading 2 "Docs" link "All docs" @22 link "Architecture" @23 link "Run record" @24 link "Telemetry" @25 link "Grafana" @26 heading 2 "Project" link "GitHub" @27 link "Contributing" @28 ...Step 4: scroll down

- Page
- How it works | Dummy Run https://dummyrun.dev/how-it-works/ · 200 · 290 ms
- Action
- scroll, target down
Step 5: scroll down

- Page
- How it works | Dummy Run https://dummyrun.dev/how-it-works/ · 200 · 300 ms
- Action
- scroll, target down
Step 6: click Architecture

- Page
- Architecture | Dummy Run https://dummyrun.dev/docs/architecture/ · 200 · 634 ms
- Persona said
- I opened Architecture and saw that the system has a runner, a reviewer, a results board, and telemetry and exporter pieces. That gave me the basic operating model.
- Action
- click, target Architecture
- Findings
- The docs say the runner writes the run record and the reviewer completes it, which is the closest thing to a verification gate
Accessibility outline the persona read
link "Skip to content" @1 banner link "Dummy Run" @2 navigation "Site" link "How it works" @3 link "Who it is for" @4 link "Persona library" @5 link "Docs" @6 [current] link "GitHub" @7 link "Get started" @8 button "Switch to light theme" @9 main navigation "Breadcrumb" link "Home" @10 link "Docs" @11 heading 1 "Architecture" heading 2 "Stages" heading 2 "Adapters" link "PROVIDERS.md" @13 heading 2 "The evidence rule" heading 2 "Browsers" heading 2 "Deployment" complementary "Next step" link "Read Run record" @15 link "docs/ARCHITECTURE.md" @16 navigation "Docs" link "Docs" @17 link "The engine" @18 link "Command line reference" @19 link "Model providers" @20 link "Personas" @21 link "Coverage" @22 link "Architecture" @23 [current] link "Run record" @24 link "Telemetry" @25 link "Grafana dashboards" @26 link "Grafana guide" @27 link "Plan" @28 link "Contributing to Dummy Run" @29 link "Writing and citation guide" @30 ...Step 7: scroll down

- Page
- Architecture | Dummy Run https://dummyrun.dev/docs/architecture/ · 200 · 355 ms
- Action
- scroll, target down
Step 8: scroll down

- Page
- Architecture | Dummy Run https://dummyrun.dev/docs/architecture/ · 200 · 290 ms
- Action
- scroll, target down
Step 9: click Read Run record

- Page
- Run record | Dummy Run https://dummyrun.dev/docs/run-record/ · 200 · 842 ms
- Action
- click, target Read Run record
Accessibility outline the persona read
link "Skip to content" @1 banner link "Dummy Run" @2 navigation "Site" link "How it works" @3 link "Who it is for" @4 link "Persona library" @5 link "Docs" @6 [current] link "GitHub" @7 link "Get started" @8 button "Switch to light theme" @9 main navigation "Breadcrumb" link "Home" @10 link "Docs" @11 heading 1 "Run record" link "telemetry" @12 heading 2 "Fields" heading 3 "Top level" heading 3 "persona" heading 3 "scenario" heading 3 "target" heading 3 "provider" heading 3 "browser" heading 3 "usage" heading 3 "persona.dimensions" heading 3 "target.route_universe, each item" heading 3 "steps, each item" heading 3 "findings, each item" heading 2 "An example record" button "Show the record (200 lines of JSON)" @24 [collapsed] complementary "Next step" link "Read Telemetry" @26 link "site/content/docs/run-record.md" @27 navigation "Docs" link "Docs" @28 link "The engine" @29 link "Command line reference" @30 link "Model providers" @31 link "Personas" @32 ...Step 10: scroll down

- Page
- Run record | Dummy Run https://dummyrun.dev/docs/run-record/ · 200 · 325 ms
- Action
- scroll, target down
Step 11: scroll down

- Page
- Run record | Dummy Run https://dummyrun.dev/docs/run-record/ · 200 · 323 ms
- Action
- scroll, target down
Step 12: scroll down

- Page
- Run record | Dummy Run https://dummyrun.dev/docs/run-record/ · 200 · 390 ms
- Action
- scroll, target down
Step 13: scroll down

- Page
- Run record | Dummy Run https://dummyrun.dev/docs/run-record/ · 200 · 305 ms
- Action
- scroll, target down
Step 14: scroll down

- Page
- Run record | Dummy Run https://dummyrun.dev/docs/run-record/ · 200 · 315 ms
- Action
- scroll, target down
Step 15: scroll down

- Page
- Run record | Dummy Run https://dummyrun.dev/docs/run-record/ · 200 · 326 ms
- Action
- scroll, target down
Step 16: scroll down

- Page
- Run record | Dummy Run https://dummyrun.dev/docs/run-record/ · 200 · 314 ms
- Action
- scroll, target down
Step 17: scroll down

- Page
- Run record | Dummy Run https://dummyrun.dev/docs/run-record/ · 200 · 369 ms
- Action
- scroll, target down
Step 18: look

- Page
- Run record | Dummy Run https://dummyrun.dev/docs/run-record/ · 200 · 280 ms
- Action
- look
Step 19: scroll down

- Page
- Run record | Dummy Run https://dummyrun.dev/docs/run-record/ · 200 · 349 ms
- Action
- scroll, target down
Step 20: scroll down

- Page
- Run record | Dummy Run https://dummyrun.dev/docs/run-record/ · 200 · 287 ms
- Action
- scroll, target down
Step 21: click Show the record (200 lines of JSON)

- Page
- Run record | Dummy Run https://dummyrun.dev/docs/run-record/ · 200 · 349 ms
- Persona said
- On Run record, I found the strongest evidence for how findings are checked: the page says the run record is the source of truth, the runner writes it, and the reviewer completes it.
- Action
- click, target Show the record (200 lines of JSON)
- Focus
- button "Show the record (200 lines of JSON)"
- Findings
- The docs say the runner writes the run record and the reviewer completes it, which is the closest thing to a verification gate
The run record example shows Firefox as a supported browser in the recorded run
I found browser details in the schema/example, but not a dedicated supported-browsers list
Accessibility outline the persona read
link "Skip to content" @1 banner link "Dummy Run" @2 navigation "Site" link "How it works" @3 link "Who it is for" @4 link "Persona library" @5 link "Docs" @6 [current] link "GitHub" @7 link "Get started" @8 button "Switch to light theme" @9 main navigation "Breadcrumb" link "Home" @10 link "Docs" @11 heading 1 "Run record" link "telemetry" @12 heading 2 "Fields" heading 3 "Top level" heading 3 "persona" heading 3 "scenario" heading 3 "target" heading 3 "provider" heading 3 "browser" heading 3 "usage" heading 3 "persona.dimensions" heading 3 "target.route_universe, each item" heading 3 "steps, each item" heading 3 "findings, each item" heading 2 "An example record" button "Show the record (200 lines of JSON)" @24 [expanded] <- focus complementary "Next step" link "Read Telemetry" @26 link "site/content/docs/run-record.md" @27 navigation "Docs" link "Docs" @28 link "The engine" @29 link "Command line reference" @30 link "Model providers" @31 link "Personas" @32 ...Step 22: scroll down

- Page
- Run record | Dummy Run https://dummyrun.dev/docs/run-record/ · 200 · 353 ms
- Action
- scroll, target down
- Focus
- button "Show the record (200 lines of JSON)"
Step 23: look

- Page
- Run record | Dummy Run https://dummyrun.dev/docs/run-record/ · 200 · 322 ms
- Persona said
- The run shown in the example uses a Firefox browser, and the fields table says browser is part of the record.
- Action
- look
- Focus
- button "Show the record (200 lines of JSON)"
- Findings
- The run record example shows Firefox as a supported browser in the recorded run
I found browser details in the schema/example, but not a dedicated supported-browsers list
Step 24: finish pass

- Page
- Run record | Dummy Run https://dummyrun.dev/docs/run-record/ · 200 · 287 ms
- Note
- verdict: pass
- Action
- finish, target pass
- Focus
- button "Show the record (200 lines of JSON)"