Phase: Harden — making the existing flow actually work
Fix the Step 3 audio polish dead end so that after a creator applies polish, the normal workspace flow offers a forward action and moves into the next production step with the same episode, speakers, selected style, social context, and the polished-track results intact. This is the next product step because the core end-to-end episode workflow currently stops after audio polish even though the creator has completed the step successfully.
Acceptance (definition of done)
In the running product, a creator can start or open an episode, reach Step 3 audio polish, choose any quality preset, apply it, see the polished tracks saved, then click a forward action that advances the step indicator to Step 4 and opens the visual moments/captions/overlays workflow for the same episode. The Step 4 screen must have access to the polished-track outputs and must still show the correct speakers and episode context; going back to setup must remain available but must not be the only post-apply path.
Maintainer review standard
The maintainer will verify this through the normal product workflow, using the running product and the repository checks as evidence. Demo-only shortcuts, seeded mock outputs, or verifier-specific paths do not count unless this step explicitly asks for them. If repeated PRs show this step is blocked or badly framed, the maintainer may close it without marking it done and open a prerequisite step instead.
This is the current active step for this product. The maintainer will only merge a pull request that completes THIS step and demonstrably works. Open a PR that closes this issue.
Why this step now: Checklist: create a new episode from Riverside/uploaded synced speaker files = works; assign files to Host/Guest buckets = works; add social links for smarter names/topics/references/spellings = works; choose preset visual style, preview it, and apply it = works; open a reusable canvas editor for layered podcast layouts = works; clean and balance episode audio with creator-facing controls = partial because polish can be applied but the workflow dead-ends immediately afterward; add contextual captions/b-roll/callouts/title moments across a full episode = partial because the feature exists but the normal flow cannot reach it from audio polish; save reusable show templates = works; export a polished long-form episode = partial because export exists but the end-to-end path is blocked before later review/export steps. Open bug #204 is a validation hardening issue and #200 is an hour-plus timestamp bug, but the maintainer assessment shows a higher-priority core-flow blocker: after applying audio polish, Continue does not continue. This step targets that dead end so the existing MVP path can actually proceed toward the rest of the promised workflow.
Phase: Harden — making the existing flow actually work
Fix the Step 3 audio polish dead end so that after a creator applies polish, the normal workspace flow offers a forward action and moves into the next production step with the same episode, speakers, selected style, social context, and the polished-track results intact. This is the next product step because the core end-to-end episode workflow currently stops after audio polish even though the creator has completed the step successfully.
Acceptance (definition of done)
In the running product, a creator can start or open an episode, reach Step 3 audio polish, choose any quality preset, apply it, see the polished tracks saved, then click a forward action that advances the step indicator to Step 4 and opens the visual moments/captions/overlays workflow for the same episode. The Step 4 screen must have access to the polished-track outputs and must still show the correct speakers and episode context; going back to setup must remain available but must not be the only post-apply path.
Maintainer review standard
The maintainer will verify this through the normal product workflow, using the running product and the repository checks as evidence. Demo-only shortcuts, seeded mock outputs, or verifier-specific paths do not count unless this step explicitly asks for them. If repeated PRs show this step is blocked or badly framed, the maintainer may close it without marking it done and open a prerequisite step instead.
This is the current active step for this product. The maintainer will only merge a pull request that completes THIS step and demonstrably works. Open a PR that closes this issue.
Why this step now: Checklist: create a new episode from Riverside/uploaded synced speaker files = works; assign files to Host/Guest buckets = works; add social links for smarter names/topics/references/spellings = works; choose preset visual style, preview it, and apply it = works; open a reusable canvas editor for layered podcast layouts = works; clean and balance episode audio with creator-facing controls = partial because polish can be applied but the workflow dead-ends immediately afterward; add contextual captions/b-roll/callouts/title moments across a full episode = partial because the feature exists but the normal flow cannot reach it from audio polish; save reusable show templates = works; export a polished long-form episode = partial because export exists but the end-to-end path is blocked before later review/export steps. Open bug #204 is a validation hardening issue and #200 is an hour-plus timestamp bug, but the maintainer assessment shows a higher-priority core-flow blocker: after applying audio polish, Continue does not continue. This step targets that dead end so the existing MVP path can actually proceed toward the rest of the promised workflow.