unpitchdocs
Troubleshooting

What the run statuses mean

Tell a queued run from a running one, a failed run from a result that will not load, and know how long to wait before you press anything again.

You pressed Run or Simulate, and what is on screen is not a result. Before pressing anything a second time, work out which of four states the run is in, because each one implies something different about your credits and about whether waiting will help.

The four states

Every AI run in unpitch is waiting to start, running, done, or failed. Only the first is printed as a word: QUEUED, on the name-plate at the top of the panel. The other three you read off the headings.

  • Waiting to start. The job exists. Nothing has read your text yet.
  • Running. Your text is being read, and the run is reporting its current phase.
  • Done. A result was saved. It may still take a moment to appear.
  • Failed. No result was saved, and nothing was charged.

Waiting to start

Quality Check shows the chip QUALITY CHECK · QUEUED with the heading Preparing your diagnostic map and says Safe to leave — this job continues in the background. Persona Reaction shows PERSONA REACTION · QUEUED with Getting the reader ready and says Safe to leave — this reaction continues in the background.

That line is literal. Closing the tab does not cancel the run. The app writes the run into the address bar as a ?run= parameter, so reloading the document reattaches you to the same run instead of starting a second one.

Running

Quality Check switches to Building your diagnostic map, starts an elapsed timer, and lands findings one at a time with a pulsing placeholder where the next will go: Findings appear as the check moves through the text.

Simulation switches to Reaction in progress and moves through three honest phases: Preparing context, Reading the draft, and Interpreting reaction. There is no clock, raw role-play prose, partial score, or partial verdict in this panel. The footer explains that the complete, structured reaction appears only after the full draft has been read and interpreted.

Two differences between the two runs matter while they are going:

  • A Simulation makes the whole document non-editable from the press of Simulate until the reaction lands. The Quality Check tab explains it with a banner: Simulation is still reading this draft. Editing is paused until it finishes. A Quality Check does not pause editing.
  • There is no way to cancel a running Simulation. Leaving the panel or pressing Run another reader only stops you watching. The run continues and the 10 credits stay spent.

A researched persona build has its own running view: the stage rows Researching segment, Authoring psychology and Review pass, with a clock on the active one and done beside each finished one.

Done

Quality Check shows Check complete for a moment, then the verdict. Persona Reaction shows Finalizing the reaction with The full read is complete. The structured reaction is being assembled., then the reaction.

Done means the result was saved, which is not the same as the result being on your screen. If the panel says Couldn’t load the findings just now. with a View results button, that is a display problem: The result is saved — this only retries loading it. Simulation's version is a card headed Simulation complete with Couldn’t load your reader’s reactions just now. and a Try again that only re-fetches.

Failed

Failed means no result was saved and your draft is untouched.

  • Quality Check: Quality Check didn’t finish, with No result was saved. Your text and the previous snapshot are unchanged. and A failed job is not charged.
  • Simulation: That simulation didn’t finish, with The previous 10-credit hold was released. Trying again starts a new Simulation for 10 credits.
  • A persona build: Research didn’t finish.

The two Try again buttons in Simulation are not the same button. On the failed-run card it starts a new run and costs another 10 credits. On the Simulation complete card it only re-fetches a result you have already paid for.

Reading it from the sidebar

While a Quality Check or a Simulation is in flight, that document's row in Projects or Recents swaps its type icon for a slowly pulsing dot. It does not say which run is going, and one document can have both, so open it to find out.

The dot refreshes about every ten seconds while the tab is visible and stops while the tab is hidden, so a dot in a background tab is not current news. Switching back to the tab refreshes it immediately.

How long to wait

Each surface states its own expectation, and none is a timeout you have to enforce yourself.

  • Simulation prints Deep read · ~1–2 min beside the Simulate button.
  • A researched persona build says it takes 3 to 5 minutes, and declares itself stuck on its own after twelve minutes without progress: This build stopped reporting progress. Start a new one — nothing was saved.
  • Quality Check gives no estimate, but it gives you the timer and a stream of findings, which is better evidence.

Judge by movement rather than by the clock. If findings are still arriving or the Simulation phase is advancing, the run is fine and long. If nothing has moved, reload the page rather than pressing the button again: the ?run= link reconnects you to the run you already have, and a second press can be a second charge.