@aicopycoders/exodus 2026.9.1900-beta.2 → 2026.9.2000-beta.1

package/assets/docs/CLAUDE.md

@@ -61,7 +61,7 @@ These are the **Exodus** pipelines — the automatic system that drives the Gene
 | **Meme** | `exodus-meme` | Meme ads — recommends formats for a brief, then one batched run renders AI-image and classic Imgflip-template memes server-side. |
 | **Workflow** | `exodus-workflow` | Run and build saved multi-node workflow automations from the CLI, resolve checkpoint/repair parks from the inbox, continue bot sessions, fire triggers, and read banks or promote winners. |
 | **Hooks** | `exodus-hooks` | Read the Scout library — the hook cards captured from Instagram outliers. Compose filters (score, lane, validation status, language, hook type, recency), inspect one card, find cards sharing a hook pattern, export to CSV. Read-only. |
-| **Video** | `exodus-video` | Make a video ad from a script. Main path: run a saved video workflow (`npx @aicopycoders/exodus workflow run`, no Show); starting from a Show (`video start --show`) is the other way in. Choose the voices as the run starts: `--voices` gives each speaker an ElevenLabs voice, `--voice-treatment` says how the voices are made (including having the video model speak a voice you wrote). Then read and approve the storyboard, `npx @aicopycoders/exodus video pull` every piece (per-scene clips, narration and per-scene voice, keyframes, cast stills, word timings, music bed) plus a manifest, cut them in whatever editor you use (a bundled script is one option), upload the cut and approve it. Admin-only. |
+| **Video** | `exodus-video` | Make a video ad from a script by running a saved video workflow (`npx @aicopycoders/exodus workflow run`). Choose the voices as the run starts: `--voices` gives each speaker an ElevenLabs voice, `--voice-treatment` says how the voices are made (including having the video model speak a voice you wrote). Then read and approve the storyboard, `npx @aicopycoders/exodus video pull` every piece (per-scene clips, narration and per-scene voice, keyframes, cast stills, word timings, music bed) plus a manifest, cut them in whatever editor you use (a bundled script is one option), upload the cut and approve it. Admin-only. |
 | **Browse** | `exodus-browse` | View history and retrieve past outputs |
 | **Drive** | `exodus-drive` | Read/write Google Docs, Sheets, and Drive files via `npx @aicopycoders/exodus drive` (uses the dashboard's Google OAuth — no local CLI) |
 | **Swipe / Mining** | (no skill; CLI only) | The competitor watch list behind the Mining page: `npx @aicopycoders/exodus swipe brands list`, `swipe brands bulk-import <csv>` (checks each brand and starts collecting its ads), `swipe brands add "<Name>" --fb <pageId>` (adds one brand and starts checking it the same way), `swipe run` / `swipe mine`. There is no `competitors` verb. |

package/assets/docs/PIPELINES.md

@@ -16,7 +16,7 @@ In Claude Code you don't click buttons. You say **"exodus"** plus what you want
 | exodus-meme | Meme Ad Generator — recommend formats, one batched server-side run |
 | exodus-workflow | Run/build saved multi-node workflows, resolve inbox gate/repair parks, continue sessions, fire triggers, read banks + promote winners |
 | exodus-hooks | Read the Scout library of hook cards — composable filters, one-card detail, pattern matches, CSV export (read-only) |
-| exodus-video | Make a video ad from a Show — start, storyboard gate, pull every piece + manifest, cut it in whatever editor you use, upload, approve (admin-only) |
+| exodus-video | Make a video ad from a script by running a saved video workflow — storyboard gate, pull every piece + manifest, cut it in whatever editor you use, upload, approve (admin-only) |
 | exodus-browse | View History and Surface the Right Run |
 | exodus-drive | Google Drive, Docs, Sheets via the Dashboard's OAuth |
 | exodus-winners | Mine your own Meta ad account for winners (needs the Meta Ads MCP) and import them into Exodus |
@@ -252,24 +252,20 @@ Returns: a table, --json data, or an exported CSV/JSON file
 
 ## exodus-video
 
-**What it does:** Makes a video ad from a script, in pieces. Main path is a saved video workflow with no Show (`exodus workflow run <name> --input <field>=@script.txt`); a locked Show (`exodus video start --show`) is the other way in and is the only route with `video flag`. On the workflow route the run stops for your cut only if its video node sets `finalWatch: true` (defaults to off — check with `exodus workflow export`, not `describe`), and video gates are approved with `exodus video approve`, not `workflow checkpoint approve`. Either way the run writes the storyboard, draws a picture per scene, records the voice, renders a clip per scene and generates a music bed, then parks for a cut. The CLI pulls the pieces to a folder with a `manifest.json`; you cut them in whatever editor you use; the upload attaches the cut for approval. Admin-only.
+**What it does:** Makes a video ad from a script, in pieces. You run a saved video workflow (`exodus workflow run <name> --input <field>=@script.txt`). The run stops for your cut only if its video node sets `finalWatch: true` (defaults to off — check with `exodus workflow export`, not `describe`), and video gates are approved with `exodus video approve`, not `workflow checkpoint approve`. The run writes the storyboard, draws a picture per scene, records the voice, renders a clip per scene and generates a music bed, then parks for a cut. The CLI pulls the pieces to a folder with a `manifest.json`; you cut them in whatever editor you use; the upload attaches the cut for approval. Admin-only.
 
-**Voices, on the workflow route.** Two separate choices, both made when the run starts and both refused before the run is created if anything is wrong, so a mistake costs nothing. `--voices @voices.json` says WHICH ElevenLabs voice each speaker gets, keyed by the names the script uses (`HOST 1`) — the same file `exodus video voices <run> --from` reads at the review, so one file per brand serves both moments. `--voice-treatment` says HOW the voices are made: a name on its own for a way of working that plays a voice you picked, or a `voice-treatment.json` shaped `{"path": "...", "describe": {"HOST 1": "Dry, tired baritone."}}` for the ways where the video model speaks the CAST's lines itself in voices you WROTE. A written voice needs one description per speaker, and it cannot be changed at the storyboard review — that is what the flag at launch is for. The two never combine: written voices plus `--voices` is refused, because those ElevenLabs voices would be paid for and never heard. A written voice covers the CAST's lines only: a narrated scene is still recorded by ElevenLabs on the member's own key, so a script with `NARRATOR:` turns still needs the key and still costs ElevenLabs money. Ways of working beyond the two shipped ones are switched on only on the test stack while they are being tried out.
+**Voices.** Two separate choices, both made when the run starts and both refused before the run is created if anything is wrong, so a mistake costs nothing. `--voices @voices.json` says WHICH ElevenLabs voice each speaker gets, keyed by the names the script uses (`HOST 1`) — the same file `exodus video voices <run> --from` reads at the review, so one file per brand serves both moments. `--voice-treatment` says HOW the voices are made: a name on its own for a way of working that plays a voice you picked, or a `voice-treatment.json` shaped `{"path": "...", "describe": {"HOST 1": "Dry, tired baritone."}}` for the ways where the video model speaks the CAST's lines itself in voices you WROTE. A written voice needs one description per speaker, and it cannot be changed at the storyboard review — that is what the `--voice-treatment` flag at launch is for. The two never combine: written voices plus `--voices` is refused, because those ElevenLabs voices would be paid for and never heard. A written voice covers the CAST's lines only: a narrated scene is still recorded by ElevenLabs on the member's own key, so a script with `NARRATOR:` turns still needs the key and still costs ElevenLabs money. Ways of working beyond the two shipped ones are switched on only on the test stack while they are being tried out.
 
 ```operator-guide
-Start (workflow route, no Show):
+Start:
   exodus workflow list / describe <name>                               find it and see what it needs
   exodus workflow run <workflowId|name> --input <field>=@script.txt [--wait]   start it; --wait stops at the storyboard gate
   exodus workflow run <workflowId|name> --input <field>=@script.txt --voices @voices.json   ... and give each speaker an ElevenLabs voice up front
   exodus workflow run <workflowId|name> --input <field>=@script.txt --voice-treatment @voice-treatment.json   ... or have the video model speak voices you WROTE
-Start (Show route):
-  exodus video shows [--json]                                          Shows and whether each is ready
-  exodus video start --show <id> --script <file> [--wait] [--json]     start an ad run; --wait stops at the storyboard gate
-Then, on either route:
+Then:
   exodus video storyboard <runId> [--json]                             the scene cards
   exodus video voices <runId> [--set <who>=<voiceId>] [--clear <who>] [--from <f>] [--json]  who speaks, with whose voice
   exodus video approve <runId> [--json]                                approve the storyboard, or the uploaded cut
-  exodus video flag <runId> --note "<what is wrong>" [--json]          send the storyboard back — SHOW RUNS ONLY
   exodus video retry-frame <runId> --node <nodeId> --scene <n> [--note "<t>"] [--json]  redo one still at the pixel gate
   exodus video status <runId> [--json]                                 the park and each scene's clip/voice/picture
   exodus video pull <runId> --out <dir> [--json]                       every piece + manifest.json

package/assets/skills/exodus-video/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: exodus-video
-description: Make a video ad with Exodus from the terminal. The main path is script-first: run a saved video workflow with `npx @aicopycoders/exodus workflow run`, no Show involved. Starting from a Show with `exodus video start --show` is the other way in. Either way you read and approve the storyboard, pull every finished piece (per-scene clips, the narration track and per-scene voice, keyframes, cast identity stills, word timings, the music bed) into a folder with `exodus video pull`, and upload a finished cut for approval. Exodus makes the pieces; the cut is made outside it, in whatever editor the user wants — this skill hands over the pieces and the facts about them, it does not prescribe an editor. Use it whenever the user has invoked Exodus and wants anything to do with a video ad, an ad run's clips or storyboard, a Show, "pull the pieces", "stitch it together", "make the cut", "upload my cut", or checking on a video run ("exodus, make a video ad from this script", "exodus, is my video run done", "exodus, pull the clips for run X and cut them", "exodus, what's wrong with scene 3", "exodus video status"). Also use it when the user ran an `npx @aicopycoders/exodus video` command or the `exodus` hub skill routed here. Video is admin-only: "video isn't enabled for this key" means either the dashboard user lacks the admin role or the run id is wrong, so check the id and then say so; do not retry. Never claim generic video-editing asks ("edit this mp4", "add captions to my reel") without Exodus context; in shared folders those belong to the user's other tools. Running or authoring workflows generally is `exodus-workflow`; static image ads are `exodus-image`; copy is `exodus-write`.
+description: Make a video ad with Exodus from the terminal. The main path is script-first: run a saved video workflow with `npx @aicopycoders/exodus workflow run`. You read and approve the storyboard, pull every finished piece (per-scene clips, the narration track and per-scene voice, keyframes, cast identity stills, word timings, the music bed) into a folder with `exodus video pull`, and upload a finished cut for approval. Exodus makes the pieces; the cut is made outside it, in whatever editor the user wants — this skill hands over the pieces and the facts about them, it does not prescribe an editor. Use it whenever the user has invoked Exodus and wants anything to do with a video ad, an ad run's clips or storyboard, "pull the pieces", "stitch it together", "make the cut", "upload my cut", or checking on a video run ("exodus, make a video ad from this script", "exodus, is my video run done", "exodus, pull the clips for run X and cut them", "exodus, what's wrong with scene 3", "exodus video status"). Also use it when the user ran an `npx @aicopycoders/exodus video` command or the `exodus` hub skill routed here. Video is admin-only: "video isn't enabled for this key" means either the dashboard user lacks the admin role or the run id is wrong, so check the id and then say so; do not retry. Never claim generic video-editing asks ("edit this mp4", "add captions to my reel") without Exodus context; in shared folders those belong to the user's other tools. Running or authoring workflows generally is `exodus-workflow`; static image ads are `exodus-image`; copy is `exodus-write`.
 ---
 
 # Video: make the pieces, cut the ad, hand it back
@@ -21,29 +21,18 @@ Your job in this skill is to drive that loop, hand over the pieces with the
 facts about them, and make whatever cut the user asked for. **How to edit is the
 user's call, not this skill's** — see "Cutting the ad".
 
-## Two routes in, one set of pieces out
+## How a run starts
 
-**The workflow route (no Show) is the main path.** The ad is a saved workflow in
-the brand, started script-first with `exodus workflow run`. This is where the
-real work happens and what to assume unless the user says otherwise.
-
-**The Show route** starts from a locked Show — its cast, rooms and voices set up
-on the dashboard beforehand — with `exodus video start --show`. It buys two
-things the workflow route does not have: `video flag` for sending a storyboard
-back with a note, and script-turn edits at the gate.
-
-Everything after the storyboard is the same on both: the same nodes render, the
-same `video status` reads the run, the same `video pull` writes the same folder,
-the same `video upload` and `approve` finish it. Where a behaviour differs, this
-skill says which route it belongs to. **Do not assume the routes are symmetric**
-— see "What each route can and cannot do".
+The ad is a saved workflow in the brand, started script-first with `exodus
+workflow run`. From there the same nodes render, `video status` reads the run,
+`video pull` writes the folder, and `video upload` and `approve` finish it.
 
 ## Before anything
 
 - `npx @aicopycoders/exodus video --help` and `workflow --help` are the
   authoritative flag lists. The `video` verb is admin-only and hidden from the
   top-level `--help`; that is expected, not a broken install.
-- **Video is admin-only on both routes**, and says so two different ways.
+- **Video is admin-only**, and says so two different ways.
   - "video isn't enabled for this key", from a `video` verb, has two causes the
     CLI cannot tell apart: the user is not an admin on this brand, or the run id
     does not exist. Check the id first; if it is right, they need the admin role.
@@ -58,17 +47,17 @@ skill says which route it belongs to. **Do not assume the routes are symmetric**
   you must pass `--duration <sec>`. Editing tools have their own requirements on
   top of that — the bundled script needs a full `ffmpeg` with `libx264` and
   exits 2 if it is missing.
-- The brand matters. A workflow and a Show both belong to one brand, so
-  `exodus workflow list` and `exodus video shows` show the active brand's only.
-  Wrong list means wrong brand; `exodus brand use <slug>`.
+- The brand matters. A workflow belongs to one brand, so `exodus workflow list`
+  shows the active brand's only. Wrong list means wrong brand; `exodus brand use
+  <slug>`.
 
 ## The loop
 
 Run each line as `npx @aicopycoders/exodus …`; `exodus` below is shorthand.
 
-**Starting on the workflow route (the main path).** The workflow is already
-saved in the brand — find it and read what it wants before you run it, because
-the script's input field is named by the workflow, not by this skill.
+**Starting a run.** The workflow is already saved in the brand — find it and
+read what it wants before you run it, because the script's input field is named
+by the workflow, not by this skill.
 
 ```operator-guide
 exodus workflow list                                      the brand's workflows
@@ -82,21 +71,13 @@ exodus workflow run <workflowId|name> --input <field>=@script.txt --voice-treatm
 exodus video status <runId>                               where it is, per scene
 ```
 
-**Starting on the Show route.**
-
-```operator-guide
-exodus video shows                                        which Shows are ready
-exodus video start --show <id> --script script.txt --wait  start, wait for the storyboard gate
-```
-
-**From the storyboard on, both routes are the same.**
+**From the storyboard on.**
 
 ```operator-guide
 exodus video storyboard <runId>                           read the scene cards
 exodus video voices <runId>                               who speaks and with whose voice
 exodus video voices <runId> --set <character>=<voiceId>   give a character a voice
 exodus video approve <runId>                              keep going
-exodus video flag <runId> --note "what's wrong"           send it back — SHOW RUNS ONLY
 exodus video retry-frame <runId> --node <nodeId> --scene <n>  redo one still at the pixel gate
 exodus video status <runId>                               where it is, per scene
 exodus video pull <runId> --out ./ad-<runId>              every piece + manifest.json
@@ -111,31 +92,32 @@ exodus video approve <runId>                              make it the ad
 Every command takes `--json`. Read `--json` when you need to decide something
 from the output; read the plain form when you are relaying it to the user.
 
-**Read a run's position with `exodus video status`, on either route.** It is the
-richest reader — every scene's clip, voice and picture, with the findings. A
-`--wait` that landed already printed the same stop block, and `workflow inbox`
-lists a parked video run (its row carries `hasShow`), but neither shows the
-per-scene detail, and the checkpoint verbs cannot resolve a video park — see
-"What each route can and cannot do".
+**Read a run's position with `exodus video status`.** It is the richest reader —
+every scene's clip, voice and picture, with the findings. A `--wait` that landed
+already printed the same stop block, and `workflow inbox` lists a parked video
+run, but neither shows the per-scene detail, and the checkpoint verbs cannot
+resolve a video park — see "What the CLI can and cannot do".
 
 **Where a run can stop.** Five stops, each with the exact next command printed
 under it:
 
-- storyboard gate: needs `approve`, or `flag` on a Show run.
+- storyboard gate: needs `approve`. When the stop block prints a `Redo one
+  frame:` line with a node id, one still can be redone first with `retry-frame`;
+  a gate parked before the stills are drawn has nothing to redraw.
 - final watch: every piece is made; needs `pull`, a cut, and `upload`. It is also
-  the one park where a finished clip can be redone, with `retry-clip`. A
-  workflow run only reaches this park if its video node sets `finalWatch: true`
-  — otherwise it skips straight to finished.
+  the one park where a finished clip can be redone, with `retry-clip`. A run
+  only reaches this park if its video node sets `finalWatch: true` — otherwise
+  it skips straight to finished.
 - paused: a builder checkpoint neither video verb resolves —
   `exodus workflow checkpoint <runId>` is the verb for that one.
 - failed: `status` shows how far it got, and a `--wait` that was running exits
   1 with the error. Tell the user; do not start another run on your own.
 - finished: pull what it made.
 
-**`--wait` stops at both video parks on either route** (#1818). `workflow run
---wait` and `video start --wait` both end at the storyboard gate and again at
-the final watch, exit 0, and close with the same `Parked:` block and next
-commands that `exodus video status` prints. Under `--json` the line carries an
+**`--wait` stops at both video parks** (#1818). `workflow run --wait` ends at
+the storyboard gate and again at the final watch, exits 0, and closes with the
+same `Parked:` block and next commands that `exodus video status` prints. Under
+`--json` the line carries an
 additive `stop` field in the same shape `video status --json` uses, so branch on
 `stop.at` rather than re-deriving the park. Every other park still prints its
 notice once and keeps polling, so a resolve from another shell carries the run
@@ -147,44 +129,31 @@ minutes with "Still running after 60 minutes" and exit 0; that is not done, so
 check `status` before treating it as finished. Do not poll `status` in a tight
 loop; a full run is several minutes.
 
-**What costs money.** Starting the run creates the storyboard. Approving the
-storyboard is the cost gate: it releases the paid rendering of every picture,
-voice and clip. `shows`, `workflow list`, `workflow describe`, `video status`,
+**What costs money.** Starting the run writes the storyboard and then draws the
+reference, cast and scene pictures, and only then pauses. Each picture is a few
+cents and each scene draws a few to choose from: about 25 cents on a small test,
+and it can approach a dollar on a longer ad. Approving releases the expensive
+part: the voices and the clips. So a run that has not paused yet is already
+spending a little, and a run still drawing pictures is healthy. `exodus workflow
+status` and `exodus video status` both print "This run will pause for your
+approval once the frames are ready" on a run that is headed for the gate. Do not
+cancel it because the pause has not arrived. `workflow list`, `workflow describe`, `video status`,
 `storyboard` and `pull` are free reads; run them as often as you like. Never
-start a second run to "retry" without the user asking; flag the storyboard (Show
-runs) or tell them what failed instead.
+start a second run to "retry" without the user asking; tell them what failed
+instead.
 
 **`--auto-approve` does not cover either video gate.** It releases Checkpoint
 boxes only (`convex/workflows.ts:7266` keys on `pauseReason === "checkpoint"`),
 and the storyboard gate and the final watch carry no pause reason. An unattended
-`workflow run --auto-approve` still stops dead at the cost gate — which is the
+`workflow run --auto-approve` still stops dead at the storyboard gate — which is the
 safe behaviour, since nothing should approve paid rendering or ship an uncut ad
 on its own.
 
-## What each route can and cannot do
-
-Verified against the routes, not assumed. Everything here is admin-gated on top.
-
-| Verb | Workflow run (no Show) | Show run |
-|---|---|---|
-| `video shows` | nothing to list | lists the brand's Shows |
-| `video start --show` | not the entry point | starts the run |
-| `workflow run` | starts the run | not the entry point |
-| `video status` | works | works |
-| `video storyboard` | works | works |
-| `video voices` (read) | works | works |
-| `video voices --set` | works | **refused** — the Show owns its voices |
-| `video approve` | works, both gates | works, both gates |
-| `video flag --note` | **refused** | works |
-| script-turn edits at the gate | **refused** | dashboard only |
-| `video retry-frame` | works | works |
-| `video retry-clip` | works | works |
-| `video revoice` | works | works |
-| `video pull` | works | works |
-| `video upload` | only if the video node sets `finalWatch: true` | works |
-| `workflow checkpoint approve` | **refused at a video gate** | **refused at a video gate** |
-
-Five consequences worth knowing before you promise a user anything:
+## What the CLI can and cannot do
+
+Verified against the CLI and the server, not assumed. Everything here is
+admin-gated on top. Five things worth knowing before you promise a user
+anything:
 
 - **A workflow only stops for your cut if its video node asks to.** `finalWatch`
   defaults to `false` (`convex/lib/workflow/catalog.ts:780`), and the run only
@@ -204,22 +173,20 @@ Five consequences worth knowing before you promise a user anything:
   `pauseReason === "checkpoint"` (`exodus/commands/workflow.ts:4487-4519`), and
   both video gates park with no pause reason at all, so those verbs refuse
   client-side with "Run … is not parked for a checkpoint approval — it is parked
-  at the cost gate (legacy)." `exodus video approve` posts straight through and
-  works on either route. `exodus workflow cancel` also works at a video gate; it
-  has no preflight.
-
-- **`flag` is Show-only and fails loudly.** `exodus video flag` on a workflow run
-  answers "Flag-with-note is only available on Video-module ads."
-  (`convex/videoModule.ts`, guarded on `!run.showId`). Script-turn edits refuse
-  the same way. There is **no** CLI path that sends a workflow run's storyboard
-  back with a note, and none that rewrites its scenes, script or prompts — the
-  whole-envelope edit mutation is wired only to the dashboard's gate screen, not
-  to any API route. The one exception is voices: `exodus video voices --set`
-  writes the cast's voice choices in place and nothing else (see "Giving the
-  characters voices"). For everything else the storyboard is approve-or-cancel
-  from the terminal: `exodus workflow cancel <runId>` and run again with a better
-  script, or edit the workflow. Do not offer `flag` to a user on the workflow
-  route.
+  at the cost gate (legacy)." `exodus video approve` posts straight through.
+  `exodus workflow cancel` also works at a video gate; it has no preflight.
+
+- **Nothing sends the storyboard back with a note.** There is **no** CLI path
+  that returns a run's storyboard for a rewrite, and none that rewrites its
+  scenes, script or prompts — the whole-envelope edit mutation is wired only to
+  the dashboard's gate screen, not to any API route. The one exception is
+  voices: `exodus video voices --set` writes the cast's voice choices in place
+  and nothing else (see "Giving the characters voices"). A single still can be
+  redrawn with `exodus video retry-frame`. For everything else the storyboard is
+  approve-or-cancel from the terminal: `exodus workflow cancel <runId>` and run
+  again with a better script, or edit the workflow.
+  The stop block a parked run prints may still offer a `Send back:` line. A
+  workflow ad refuses it, so never run it and never offer it to the user.
 - **`--wait` lands on both video parks, but only those.** Since #1818 every
   `--wait` caller — `workflow run`, `triggers fire`, `checkpoint approve|retry`,
   `repair retry|skip` — ends at a storyboard gate or a final watch, exits 0, and
@@ -228,20 +195,16 @@ Five consequences worth knowing before you promise a user anything:
   slots park is still not a landing: it prints its notice once and the loop
   keeps polling, so a resolve from another shell carries the run on. Do not read
   a park banner mid-poll as the wait having finished.
-- **The run's page link differs, and every link the CLI prints follows the run.**
-  A Show ad opens on the `/video` page. A showless workflow run opens on
-  `/workflows/<workflowId>/runs/<runId>`, because `/video` answers "That ad isn't
-  here" for a run with no Show (`getAdDetail` rejects it;
-  `convex/http.ts:9204-9211`). Since #1851 the manifest's `dashboardUrl`, the
-  link in every park block and the link `upload` prints are all minted that way,
-  so they agree on both routes. When all the CLI holds is a run id it mints
-  `/runs/<runId>`, the canonical forwarder, which lands on the same page.
+- **Every link the CLI prints follows the run.** A run opens on
+  `/workflows/<workflowId>/runs/<runId>`. Since #1851 the manifest's
+  `dashboardUrl`, the link in every park block and the link `upload` prints are
+  all minted that way, so they agree. When all the CLI holds is a run id it
+  mints `/runs/<runId>`, the canonical forwarder, which lands on the same page.
 
 ## The script file
 
 Plain text. A `CAST:` block naming the speakers, one line per turn, then a
-`CTA:` block. On a Show run the names must match what the Show locked. The
-planner sizes every line at
+`CTA:` block. The planner sizes every line at
 2.5 words per second (`WORD_LADDER_WPS`), so a turn near 14 words plans a clip
 near 6 seconds. If a line does get truncated in its clip, the manifest says so
 (`speech-cutoff`).
@@ -267,24 +230,23 @@ in `manifest.json` under `provenance`.
 
 ## Reading the storyboard gate
 
-`storyboard <runId>` prints one card per scene on both routes: who speaks, the
+`storyboard <runId>` prints one card per scene: who speaks, the
 line, the planned duration, and whether its picture is drawn. Under the cards it
 also lists the voices — one line per character who speaks or already has a voice
 — so the user can see who sounds like whom before paying for the clips. Read it
 for the user, then ask one question.
 
-On a **workflow run** the question is approve or don't — there is no flag. If
-the storyboard is wrong, say so plainly and let the user decide between
-approving anyway, cancelling (`exodus workflow cancel <runId>`) and re-running
-with a better script, or changing the workflow. Approving is the cost gate, so
-do not approve a storyboard you have just told the user is wrong.
+The question is approve or don't. If the storyboard is wrong, say so plainly and
+let the user decide between approving anyway, redrawing a single still when
+the stop block offers it (`exodus video retry-frame <runId> --node <nodeId>
+--scene <n>`, with the node id that block prints), cancelling
+(`exodus workflow cancel <runId>`) and re-running with a better script, or
+changing the workflow. Approving is what releases the voices and the clips — the
+expensive part — so do not approve a storyboard you have just told the user is
+wrong.
 
-On a **Show run** you can also send it back: `video flag <runId> --note "…"`.
-The note is read by the model that rewrites the storyboard, so make it concrete
-("scene 2 should be a close-up of the softgel, not the eyes").
-
-Either way, do not approve on the user's behalf unless they told you to run the
-whole loop unattended.
+Do not approve on the user's behalf unless they told you to run the whole loop
+unattended.
 
 ## Giving the characters voices
 
@@ -367,7 +329,7 @@ nobody knows the speaker names yet. When the storyboard arrives, the voices are
 already set and `exodus video voices <runId>` shows them.
 
 **Choosing how the voices are made.** `--voice-treatment` on `workflow run` says
-HOW a run makes its voices, and it is the only door on the workflow route. Two
+HOW a run makes its voices, and it is the only door. Two
 kinds of answer.
 
 A name on its own picks a way of working that plays a voice you already have,
@@ -440,8 +402,7 @@ person made at a keyboard for one run, and nobody made it for the sub-workflow.
 **Written voices cannot be changed at the review.** They are chosen when the run
 starts. `exodus video voices <runId>` on a described run names the treatment and
 prints each written voice, and says plainly that `--voice-treatment` at launch is
-what changes them. Do not tell anyone to create a Show for this, and do not set a
-voice path in node config — neither is a door.
+what changes them. Do not set a voice path in node config — it is not a door.
 
 **Where it shows up afterwards.** The launch receipt and `exodus workflow status`
 both print one line naming the treatment and the speakers who got a written
@@ -457,7 +418,10 @@ changed[]             the characters whose voice actually moved (set only)
 canChange             false once the run has left the storyboard review
 whyNot                why, in one sentence, when canChange is false
 treatment.path        which way of making voices this run froze, or null
-treatment.kind        "performance-then-conversion" or "render-owns-voice"
+treatment.kind        "performance-then-conversion", "voiceover-then-lipsync" or
+                      "render-owns-voice"
+treatment.pinsHeard   true when the chosen ElevenLabs voices are heard on this
+                      run; false only for "render-owns-voice"
 cast[]                one per character
   characterId, name
   voice               { voiceId, label }: the voice this character's clips are
@@ -514,9 +478,7 @@ manifest.json            the index below
 
 ```
 runId, pulledAt, dashboardUrl        the run and the page that opens it:
-                                     /video?ad=<runId> for a Show ad,
-                                     /workflows/<workflowId>/runs/<runId> for a
-                                     workflow run
+                                     /workflows/<workflowId>/runs/<runId>
 storyboard, reference, music         filenames, or null when not delivered
 narration                            { file, timing }, or null on a per-scene run
 cast[]                               one per identity still
@@ -554,8 +516,7 @@ provenance                           which saved rulebook this run followed:
                                      {format: {via, nodeId, rigId, rigName,
                                      rulesFrom, specVersion}}. Absent when the
                                      run followed none. `nodeId` is the Rig box
-                                     that chose it (absent on a Show ad, which
-                                     has no box); `rulesFrom` names the rig the
+                                     that chose it; `rulesFrom` names the rig the
                                      rules were borrowed from when the picked
                                      rig is a copy. Names, ids and a version
                                      only — never the rules themselves
@@ -578,12 +539,10 @@ it is a picture-only takeover of the frame while the spoken track underneath
 keeps running, so it adds no audio and takes no place in the running order. The
 envelope also carries `overlays[]` (each with a `cueLineId` and a `placement`)
 and `audio` (`roomTone`, `music`, `sfxCues[]`) — the planner's notes on what
-should sit on top. **Route difference:** on a Show run the image and motion
-prompts (`framePrompt`, `videoPrompt`, `referencePrompt`) are stripped before
-the file reaches you, for everyone including admins — they are style-pack IP
-(`convex/workflows.ts:13847`, `isModuleRun = !!run.showId`). A workflow run has
-no `showId`, so an admin key gets them in full. Their absence on a Show run is
-by design, not a broken pull.
+should sit on top. The image and motion prompts (`framePrompt`, `videoPrompt`,
+`referencePrompt`) reach an admin key in full. A caller who cannot use video on
+this brand gets them stripped before the file arrives — they are style-pack IP
+(`redactVideoIp` in `convex/workflows.ts`).
 
 **The narration track** (`narration.mp3` + `narration.json`). Present when the
 run recorded continuously: every narrated scene's `voText` joined into one
@@ -595,8 +554,7 @@ narration.mp3**. Runs go per-scene instead when any scene has on-camera
 dialogue, when the voiceover node is configured `take: "per-scene"`, or on an
 older storyboard carrying no `script[]`; then there is no `narration.mp3` and
 each scene's voice is its own recording. That choice is made from the storyboard
-and the node config (`planNarration`), not from the route — a workflow run and a
-Show run reach it the same way.
+and the node config (`planNarration`).
 
 **The per-scene voice tracks** (`scene-NN.voice.<ext>`). On a continuous run
 these are cut out of `narration.mp3` itself — the same audio, sliced at the
@@ -611,9 +569,8 @@ speech inside the clip — under the default voice treatment the video model's
 invented voice is then replaced, in place, with the cast member's pinned
 ElevenLabs voice (that is `revoiced: true`). Other treatments leave the render
 owning its own audio, e.g. lip-sync retargeting a VO track onto the video. The
-treatment is frozen on the run when it starts, and there are exactly two doors
-that set one: `--voice-path` on `video start` for a Show run, and
-`--voice-treatment` on `workflow run` for a workflow run (see **Choosing how the
+treatment is frozen on the run when it starts, and there is exactly one door
+that sets it: `--voice-treatment` on `workflow run` (see **Choosing how the
 voices are made**). Node config is never a door — nothing on the canvas sets a
 voice treatment, and a run that was given none renders the default. Read
 `wordsFrom` and `revoiced` per scene rather than assuming.
@@ -651,8 +608,8 @@ finding says `wrong-character`.
 generated by Eleven Music after the last clip lands, at the landed clips' summed
 length plus a 3-second safety pad. One track per ad, never per-scene. It is
 never mixed into any clip and never levelled — gain, fade and ducking are the
-cut-maker's decisions. `--no-music` on `start` means no artifact at all, and a
-music failure degrades to a warning rather than blocking the run.
+cut-maker's decisions. A music failure degrades to a warning rather than
+blocking the run.
 
 ### Reading it before you cut
 
@@ -715,7 +672,7 @@ revoiced and trimmed the same way, with fresh findings and fresh word timings.
 scene's motion prompt only (`withRedoDirection`,
 `scout/src/trigger/workflow-scene-retry.ts`); the spoken line still comes from
 the approved storyboard. Asking for different words in a note will not change
-them. If the words are wrong, that is a storyboard flag or a new run.
+them. If the words are wrong, that is a new run.
 
 **A redo whose take fails the checks keeps the original clip.** On a finished
 clip the row stays `done` with the ORIGINAL take and only its findings are
@@ -795,8 +752,7 @@ they want; if they have not said and are not around, say which you used and why.
 - **The bundled script** — `node .claude/skills/exodus-video/scripts/first-cut.mjs
   ./ad-<runId>`, path from the workspace root (from a brand subfolder, prefix
   `../`). A plain assembly with no editing decisions in it, for seeing the whole
-  ad end to end quickly. Read "What the bundled script does" before you run it:
-  it changes the audio.
+  ad end to end quickly. Read "What the bundled script does" before you run it.
 - **HyperFrames** — HTML/CSS compositions rendered to video. Suits an ad that
   needs captions, title cards, overlays or motion graphics laid over the clips.
   It is a separate skill pack; check that a `hyperframes` skill is installed
@@ -829,17 +785,15 @@ Read these out of `storyboard.json` before you build anything, in any tool:
 
 ### What the bundled script does
 
-State this to the user before you run it, because an editor does not expect a
-cut tool to alter the performance:
-
-- **It changes the narration's speed.** When a narrated scene's voice runs
-  longer than its clip, it speeds that scene's voice up — to at most 1.26×
-  (`ATEMPO_CAP`, `scripts/first-cut.mjs:39`) — and holds the last frame for the
-  remainder. The amount differs per scene, so one ad can play scene 1 at normal
-  speed and scene 3 a quarter faster. **This is the script's behaviour, not the
-  system's**; nothing in the pipeline does it. It is the defect tracked in
-  #1814. If the user cares about a single consistent narrator pace, cut
-  elsewhere, or skip the scenes it stretches.
+State this to the user before you run it, so they know what the cut does with
+the pieces:
+
+- **It cuts the picture to the voice, and never the other way round.** A
+  narrated scene's segment runs exactly as long as that scene's voice track plus
+  a beat of air. A clip that runs longer is trimmed; a clip that runs shorter
+  holds its last frame for the difference. The narration is never sped up,
+  slowed down or cut short, so every scene plays at the speed it was recorded,
+  and the word times are handed on as they were measured.
 - **It ignores the continuous master.** It reads only `scene-NN.voice` files;
   it never opens `narration.mp3` or `manifest.narration`. A tool that lays the
   master down whole and cuts the pictures to it keeps the read intact.
@@ -856,7 +810,8 @@ cut tool to alter the performance:
 - Beside the MP4 it writes `cut.words.json` (every spoken word of the finished
   ad with its start and end) and `cut.srt` (the same words grouped into subtitle
   lines), named from the `--out` basename. Those times are on the finished ad's
-  timeline and include the speed changes above. Scenes with no word times leave
+  timeline, each scene's own word times offset by where that scene starts in the
+  cut and otherwise unaltered. Scenes with no word times leave
   gaps and are named in the report; when no scene delivered any, neither file is
   written. These sidecars are the script's own output — another tool will not
   produce them.
@@ -880,8 +835,8 @@ give a real one.
 
 **`upload` only works at the final watch.** The run has to be parked there with
 its video step done, or the server answers "This ad isn't waiting for your final
-cut right now." On a workflow run that park exists only when the video node sets
-`finalWatch: true`; a Show run always has it. Uploading again replaces the
+cut right now." That park exists only when the video node sets
+`finalWatch: true`. Uploading again replaces the
 previous cut in place rather than adding a second one. Upload does not move the
 run; it stays parked until someone approves.
 
@@ -896,7 +851,7 @@ compares the cut against the script or the storyboard. The approved spoken copy
 and the approved narration are therefore not the editor's to change — reordering
 lines, re-recording narration, or cutting words out of the CTA will ship,
 silently. Cut the pictures to the words, not the words to the pictures. If the
-words are wrong, that is a storyboard flag or a new run, not an edit.
+words are wrong, that is a new run, not an edit.
 
 ## Reporting to the user
 

package/assets/skills/exodus-video/scripts/first-cut.mjs

@@ -9,8 +9,8 @@
 // The audio track comes first. A narrated scene (the storyboard gives it
 // narration and no on-camera line) plays its voice track and NOTHING the clip
 // recorded: the voice replaces the clip's audio, and the picture is fitted to
-// the voice, trimmed when the clip runs longer, sped up to at most 1.26x and
-// then held on its last frame when the voice runs longer. A dialogue scene
+// the voice, trimmed when the clip runs longer and held on its last frame when
+// the voice runs longer. The voice itself is never altered. A dialogue scene
 // keeps the audio it performed. A scene with no clip but a keyframe becomes a
 // still for the length of its voice track, else the storyboard's planned
 // duration. Every segment is loudness-normalized to -16 LUFS before the join.
@@ -35,8 +35,6 @@ const stereo = "aresample=48000,aformat=channel_layouts=stereo";
 const LOUDNORM = "loudnorm=I=-16:TP=-1.5:LRA=11,aresample=48000";
 /** Air after the last spoken word of a narrated scene, so the line lands. */
 export const TAIL_AIR_SEC = 0.8;
-/** The most a voice track may be sped up to fit its clip. */
-export const ATEMPO_CAP = 1.26;
 /** A continuous low bed under the whole ad, so the joins between clips do not read as dead air. */
 const ROOM_TONE =
   "anoisesrc=color=brown:amplitude=0.0035:r=48000,highpass=f=60,lowpass=f=900,aformat=channel_layouts=stereo";
@@ -91,27 +89,18 @@ export function scenePlans(storyboard) {
 }
 
 /**
- * Fit a picture to its voice track. The segment runs as long as the voice plus
- * a beat of air. A clip that runs longer is trimmed; a voice that runs longer
- * is sped up to at most ATEMPO_CAP and the clip's last frame is held for
- * whatever is still left. A still (no clip) simply runs the voice's length.
+ * Fit a picture to its voice track. The segment always runs as long as the
+ * voice plus a beat of air, because the voice is never altered. A clip that
+ * runs longer is trimmed; a clip that runs shorter holds its last frame for the
+ * difference. A still (no clip) simply runs the voice's length.
  */
 export function fitNarration({ clipSeconds, voiceSeconds }) {
   const target = voiceSeconds + TAIL_AIR_SEC;
-  if (clipSeconds === null) return { seconds: target, atempo: 1, holdSec: 0, trimmedBy: 0 };
+  if (clipSeconds === null) return { seconds: target, holdSec: 0, trimmedBy: 0 };
   if (target <= clipSeconds) {
-    return { seconds: target, atempo: 1, holdSec: 0, trimmedBy: clipSeconds - target };
+    return { seconds: target, holdSec: 0, trimmedBy: clipSeconds - target };
   }
-  const room = Math.max(clipSeconds - TAIL_AIR_SEC, 0.1);
-  const atempo = Math.min(ATEMPO_CAP, voiceSeconds / room);
-  const fitted = voiceSeconds / atempo + TAIL_AIR_SEC;
-  if (fitted <= clipSeconds) return { seconds: fitted, atempo, holdSec: 0, trimmedBy: 0 };
-  return { seconds: fitted, atempo, holdSec: fitted - clipSeconds, trimmedBy: 0 };
-}
-
-function scaleWords(words, atempo) {
-  if (!words || atempo === 1) return words;
-  return words.map((w) => ({ ...w, s: w.s / atempo, e: w.e / atempo }));
+  return { seconds: target, holdSec: target - clipSeconds, trimmedBy: 0 };
 }
 
 export function cueOffsetOnSpine(spine, counts, cueLineId) {
@@ -186,27 +175,26 @@ export function buildTimeline({ manifest, storyboard, skip, probe, exists, readW
         const fit = fitNarration({ clipSeconds: clip.seconds, voiceSeconds: probe(voice).seconds });
         const how = [];
         if (fit.trimmedBy > 0.05) how.push(`clip trimmed ${clip.seconds.toFixed(1)}s -> ${fit.seconds.toFixed(1)}s`);
-        if (fit.atempo > 1.001) how.push(`voice sped up ${fit.atempo.toFixed(2)}x`);
         if (fit.holdSec > 0.05) how.push(`last frame held ${fit.holdSec.toFixed(1)}s`);
         spine.push({
-          ...base, audio: "narration", voice, seconds: fit.seconds, atempo: fit.atempo,
-          holdSec: fit.holdSec, words: scaleWords(readWords(scene.words), fit.atempo),
+          ...base, audio: "narration", voice, seconds: fit.seconds,
+          holdSec: fit.holdSec, words: readWords(scene.words),
           note: `${flag}, ${voice} replaces the clip's audio${how.length ? ` (${how.join(", ")})` : ""}`,
         });
       } else if (narrated) {
         spine.push({
-          ...base, audio: "silence", voice: null, seconds: clip.seconds, atempo: 1, holdSec: 0,
+          ...base, audio: "silence", voice: null, seconds: clip.seconds, holdSec: 0,
           words: null,
           note: `${flag}, narrated scene without its voice track: the clip's own audio is muted`,
         });
       } else if (clip.hasAudio) {
         spine.push({
-          ...base, audio: "clip", voice: null, seconds: clip.seconds, atempo: 1, holdSec: 0,
+          ...base, audio: "clip", voice: null, seconds: clip.seconds, holdSec: 0,
           words: readWords(scene.words), note: `${flag}, its own dialogue`,
         });
       } else {
         spine.push({
-          ...base, audio: "silence", voice: null, seconds: clip.seconds, atempo: 1, holdSec: 0,
+          ...base, audio: "silence", voice: null, seconds: clip.seconds, holdSec: 0,
           words: null, note: `${flag}, silent clip`,
         });
       }
@@ -216,10 +204,10 @@ export function buildTimeline({ manifest, storyboard, skip, probe, exists, readW
     if (exists(scene.keyframe)) {
       const fit = voice
         ? fitNarration({ clipSeconds: null, voiceSeconds: probe(voice).seconds })
-        : { seconds: plan.durationSec ?? 4, atempo: 1, holdSec: 0 };
+        : { seconds: plan.durationSec ?? 4, holdSec: 0 };
       spine.push({
         sceneIndex: n, source: "still", file: scene.keyframe, clipSeconds: null,
-        seconds: fit.seconds, audio: voice ? "narration" : "silence", voice, atempo: 1, holdSec: 0,
+        seconds: fit.seconds, audio: voice ? "narration" : "silence", voice, holdSec: 0,
         startSec, lineIds: plan.lineIds,
         // #1689: on a narrated scene the words are the voice track's, and the
         // voice starts where the segment does, so they need no offset here.
@@ -288,8 +276,7 @@ export function buildFfmpegArgs(timeline, { resolve, music, out }) {
     filters.push(`${norm(`${i}:v`)}${fit.length ? `,${fit.join(",")}` : ""}[v${i}]`);
     if (seg.audio === "narration") {
       const v = addInput("-i", resolve(seg.voice));
-      const tempo = seg.atempo > 1.001 ? `atempo=${seg.atempo.toFixed(4)},` : "";
-      filters.push(`[${v}:a]${stereo},${tempo}${LOUDNORM},apad,atrim=0:${sec(seg.seconds)}[a${i}]`);
+      filters.push(`[${v}:a]${stereo},${LOUDNORM},apad,atrim=0:${sec(seg.seconds)}[a${i}]`);
     } else if (seg.audio === "clip") {
       filters.push(`[${i}:a]${stereo},${LOUDNORM},apad,atrim=0:${sec(seg.seconds)}[a${i}]`);
     } else {

package/assets/skills/exodus-workflow/SKILL.md

@@ -237,7 +237,7 @@ Two limits worth knowing before you treat an empty inbox as proof of anything:
 
 **What `inbox` does not list is work that isn't waiting on anybody.** A queued run, a running run, and any automation that ran start-to-finish without an approval step never appear there — not because they're hidden, but because there is nothing for you to decide. Plenty of workflows are built that way on purpose; the shipped **Winner Flywheel** starter is one (brief → splitter → output, no Checkpoint box anywhere), so its run *can never* show up in `inbox`. An empty inbox after a promote means "nothing needs you", not "nothing ran".
 
-**To see runs that aren't waiting on you, run `workflow status` with no `--id`.** That prints the brand's recent runs — newest first, one row each with the workflow name, its status, the **date** it started, and its **run id**. That bare list is how you discover a background run at all, including one that started and finished while you weren't looking; take the id from it and `status --id <runId>` for the full detail.
+**To see runs that aren't waiting on you, run `workflow status` with no `--id`.** That prints the brand's recent runs — newest first, one row each with the run's own name (the name you gave it, else the first line of its brief, else the workflow's name), its type ("Copy run" or "Workflow"), its status, the **date** it started, and its **run id**. That bare list is how you discover a background run at all, including one that started and finished while you weren't looking; take the id from it and `status --id <runId>` for the full detail.
 
 Two things about that list you have to know before you answer a question with it:
 

package/dist/commands/video.d.ts

@@ -1,6 +1,6 @@
 import { type ApiResponse } from "../lib/client.js";
 import type { FlagOccurrence } from "../lib/args.js";
-import { type RunProvenance } from "./workflow.js";
+import { type RunProvenance, type WorkflowRun } from "./workflow.js";
 export declare const helpText: string;
 export declare const VOICES_PATH = "/api/v2/video/voices";
 export interface ClipWord {
@@ -100,6 +100,7 @@ export interface VideoRun {
     pauseReason?: BuilderPauseReason;
     pausedNodeId?: string;
     nodes: VideoRunNode[];
+    pauseAhead?: WorkflowRun["pauseAhead"];
     castLock?: PullCastLock | null;
     provenance?: RunProvenance;
     workflowId?: string;
@@ -115,6 +116,7 @@ export interface NodeItem {
     attempt?: number;
     flagged?: boolean;
     findings?: ClipFinding[];
+    staleClaim?: boolean;
     artifact?: ArtifactSubset;
 }
 export interface ShowRow {
@@ -296,6 +298,7 @@ export interface ClipRedoTarget {
 }
 export declare function planClipRedo(run: VideoRun, items: NodeItem[], target: ClipRedoTarget): ClipRedoPlan;
 export declare function retryClipFlow(runId: string, target: ClipRedoTarget, note: string | undefined, json: boolean, deps: VideoDeps): Promise<FlowResult>;
+export declare const VOICE_PATHS_NEVER_HEARD: Set<string>;
 export declare function planClipRevoice(run: VideoRun, items: NodeItem[], target: ClipRedoTarget): ClipRedoPlan;
 export type RevoiceTarget = ClipRedoTarget | {
     all: true;

package/dist/commands/video.js

@@ -5,9 +5,9 @@ import { apiGet, apiPost, getDashboardUrl } from "../lib/client.js";
 import { displayRunStatus, formatApiError } from "../lib/format.js";
 import { missingRouteLine } from "../lib/route-support.js";
 import { hasBinary } from "../lib/preflight.js";
-import { ASSET_UPLOAD_POLICY } from "./workflow.js";
+import { ASSET_UPLOAD_POLICY, pauseAheadLine, } from "./workflow.js";
 export const helpText = `
-exodus video — make an ad from a Show, pull every piece, upload your cut
+exodus video — make a video ad from a saved workflow, pull every piece, upload your cut
 
 The dashboard makes the PIECES of an ad: the storyboard, one picture per scene,
 one voice track per scene, one video clip per scene, plus the music bed. It does
@@ -16,28 +16,34 @@ with whatever editing tools you like, and upload the cut back.
 
 The whole loop, in order:
 
-  1. exodus video shows
-     Which Shows exist and which are ready to make ads.
+  1. exodus workflow list
+     Which workflows this brand has saved. A video workflow is one that makes
+     ads.
 
-  2. exodus video start --show <id> --script script.txt --wait
-     Starts an ad and waits. It stops when the storyboard needs your yes.
+  2. exodus workflow describe <workflowId|name>
+     What that workflow needs before it can run, and what it gives back.
 
-  3. exodus video storyboard <runId>
+  3. exodus workflow run <workflowId|name> --input <field>=@script.txt --wait
+     Starts the ad on your script and waits. It stops when the storyboard
+     needs your yes. "describe" names the field your script goes in. You can
+     also settle the voices here, at the launch, instead of at step 5 — the
+     launch flags all live in: exodus workflow --help
+
+  4. exodus video storyboard <runId>
      The scene cards: what each scene says and the picture it will look like.
 
-  4. exodus video voices <runId>
+  5. exodus video voices <runId>
      Who is in the ad, which voice each one has, how the voices get applied
      and who pays. Add --set to give someone a voice, before any clip is made.
 
-  5. exodus video approve <runId>          (looks right — keep going)
-     exodus video flag <runId> --note "…"  (something is wrong — send it back)
+  6. exodus video approve <runId>          (looks right — keep going)
      exodus video retry-frame <runId> --node <nodeId> --scene <n>
        Redo ONE still at the pixel gate. Neighbours stay. The gate holds.
 
-  6. exodus video status <runId>
+  7. exodus video status <runId>
      Where the run is and how each scene's clip turned out.
 
-  7. exodus video pull <runId> --out ./ad
+  8. exodus video pull <runId> --out ./ad
      Writes every piece to that folder plus a manifest.json index.
 
      exodus video retry-clip <runId> --scene <n>
@@ -50,21 +56,18 @@ The whole loop, in order:
        picture stays, no new video is made, and a clip whose voice pass
        cannot finish is kept as it was. Use --scene <n> for one clip.
 
-  8. Make your cut from those files.
+  9. Make your cut from those files.
 
-  9. exodus video upload <runId> --file cut.mp4
-     Attaches your cut to the run and prints the page to approve it on.
+  10. exodus video upload <runId> --file cut.mp4
+      Attaches your cut to the run and prints the page to approve it on.
 
-  10. exodus video approve <runId>
-      Or click Approve on the page from step 9.
+  11. exodus video approve <runId>
+      Or click Approve on the page from step 10.
 
 Usage:
-  exodus video shows [--json]
-  exodus video start --show <id> --script <file> [--voice-path <v>] [--no-music] [--wait] [--json]
   exodus video status <runId> [--json]
   exodus video storyboard <runId> [--json]
   exodus video approve <runId> [--json]
-  exodus video flag <runId> --note "<what is wrong>" [--json]
   exodus video retry-frame <runId> --node <nodeId> --scene <n> [--note "..."] [--json]
   exodus video retry-clip <runId> --scene <n> [--node <nodeId>] [--note "..."] [--json]
   exodus video revoice <runId> (--scene <n> | --all) [--node <nodeId>] [--json]
@@ -73,18 +76,11 @@ Usage:
   exodus video upload <runId> --file <cut.mp4> [--duration <sec>] [--json]
 
 Options:
-  --show <id>          Which Show the ad belongs to (from: exodus video shows)
-  --script <file>      A text file with the words you want the ad to say
-  --voice-path <v>     Override the Show's voice path for this ad
-  --no-music           Skip the music bed
-  --wait               Stay open and report as the run moves; stops at every
-                       point that needs you
   --out <dir>          Folder to write the pulled pieces into (pull)
   --file <cut.mp4>     Your finished cut (upload). MP4, MOV or WebM, up to 200MB
   --duration <sec>     How long your cut is, in seconds. Only needed when the
                        length can't be read off the file itself
-  --note "<text>"      What is wrong with the storyboard (flag), or how to
-                       steer one redo (retry-frame, retry-clip)
+  --note "<text>"      How to steer one redo (retry-frame, retry-clip)
   --node <nodeId>      Which scene-frames node holds the still (retry-frame).
                        Which video step holds the clip, when a scene has one on
                        more than one step (retry-clip, revoice)
@@ -99,14 +95,15 @@ Options:
   --from <file.json>   A file of characters and voice IDs (voices). --set and
                        --clear win over the same name in the file
   --json               Machine-readable output
-  --help, -h           Show this help
+  --help, -h           Print this help
 
 Video is admin-only. If every command here answers "video isn't enabled for
 this key", your dashboard user needs the admin role on this brand.
 
 Examples:
-  exodus video shows
-  exodus video start --show k57abc --script ./script.txt --wait
+  exodus workflow list
+  exodus workflow describe "Video Ad"
+  exodus workflow run "Video Ad" --input script=@./script.txt --wait
   exodus video storyboard run_123
   exodus video approve run_123
   exodus video retry-frame run_123 --node frames-1 --scene 2
@@ -921,6 +918,8 @@ const ITEM_STATUS_WORD = {
 function itemWord(item) {
     if (!item)
         return "—";
+    if (item.staleClaim === true)
+        return "stopped";
     if (item.flagged === true)
         return "flagged";
     return ITEM_STATUS_WORD[item.status] ?? item.status;
@@ -974,6 +973,7 @@ export async function statusFlow(runId, json, deps) {
     const lines = [
         headline,
         ...stopLines(stop, runId, reviewUrl(deps.dashboardUrl, run)),
+        ...(stop.at === "running" && run.pauseAhead ? [pauseAheadLine(run.pauseAhead)] : []),
         ...warnings.map((w) => `Heads-up (${w.step}): ${w.warning}`),
     ];
     if (byScene.size === 0) {
@@ -1176,6 +1176,7 @@ export function planClipRedo(run, items, target) {
         return { ok: false, reason: found.reason };
     const { row, uploadedCut } = found;
     const redoable = row.status === "failed" ||
+        row.staleClaim === true ||
         (row.status === "done" && (row.flagged === true || stop.at === "final-watch"));
     if (!redoable) {
         if (row.status === "done") {
@@ -1228,7 +1229,7 @@ function findClipRow(run, items, target) {
                 "and one clip can only be redone on a step that finished.",
         };
     }
-    if (row.status === "running") {
+    if (row.status === "running" && row.staleClaim !== true) {
         return { reason: `Scene ${target.sceneIndex} is already being redone.` };
     }
     return { row, uploadedCut };
@@ -1294,20 +1295,34 @@ export async function retryClipFlow(runId, target, note, json, deps) {
 }
 function clipIsStillRaw(item) {
     const a = item.artifact;
-    return (item.status === "done" &&
+    return ((item.status === "done" || item.staleClaim === true) &&
         a?.type === "video" &&
         a.revoiced !== true &&
         a.speechTrimmed !== true &&
         a.tailTrimmed !== true);
 }
 const VOICE_NOT_CHANGED_CODES = new Set(["voice-unpinned", "voice-not-applied"]);
+export const VOICE_PATHS_NEVER_HEARD = new Set(["native-prompt", "omni-audio-ids", "gemini-direct"]);
+async function neverHeardRunLine(runId, deps) {
+    const res = await deps
+        .get(`${VOICES_PATH}?runId=${encodeURIComponent(runId)}`)
+        .catch(() => null);
+    if (!res?.ok)
+        return null;
+    const path = res.data?.treatment?.path;
+    if (typeof path !== "string" || !VOICE_PATHS_NEVER_HEARD.has(path))
+        return null;
+    return (`This run's clips are not voiced by ElevenLabs — the "${path}" way settles each voice as ` +
+        "the clip is rendered — so no clip on this run has a voice pass to redo. Nothing was queued " +
+        "and nothing was spent. To make a clip again from scratch, redo the clip instead.");
+}
 export function planClipRevoice(run, items, target) {
     const found = findClipRow(run, items, target);
     if ("reason" in found)
         return { ok: false, reason: found.reason };
     const { row, uploadedCut } = found;
     const remake = `To make the clip again from scratch: exodus video retry-clip ${run._id} --scene ${target.sceneIndex}`;
-    if (row.status !== "done" || row.artifact?.type !== "video") {
+    if ((row.status !== "done" && row.staleClaim !== true) || row.artifact?.type !== "video") {
         return {
             ok: false,
             reason: `Scene ${target.sceneIndex} has no finished clip, so there is no voice to redo. ${remake}`,
@@ -1351,10 +1366,11 @@ export async function revoiceFlow(runId, target, json, deps) {
             .map((i) => ({ sceneIndex: i.sceneIndex, nodeId: i.nodeId }))
         : [target];
     if (targets.length === 0) {
-        const nothing = "No clip on this run is waiting for a voice pass, so nothing was started.";
+        const note = (await neverHeardRunLine(runId, deps)) ??
+            "No clip on this run is waiting for a voice pass, so nothing was started.";
         return {
             code: 0,
-            lines: json ? [JSON.stringify({ ok: true, runId, scenes: [], note: nothing })] : [nothing],
+            lines: json ? [JSON.stringify({ ok: true, runId, scenes: [], note })] : [note],
         };
     }
     const scenes = [];
@@ -1551,7 +1567,8 @@ export function voiceSheetLines(sheet) {
     for (const row of sheet.cast) {
         const voice = voiceWords(row);
         lines.push(`  ${row.characterId}  ${row.name}  ${voice}  ${availabilityWords(row)}  ` +
-            `speaks in ${row.spokenScenes} scenes, about ${row.spokenSeconds}s`);
+            `speaks in ${row.spokenScenes} scene${row.spokenScenes === 1 ? "" : "s"}, ` +
+            `about ${row.spokenSeconds}s`);
         if (row.description)
             lines.push(`      written voice: ${row.description}`);
     }
@@ -1565,8 +1582,9 @@ export function voiceSheetLines(sheet) {
             : "Nothing changed — those voices were already set. Pictures and script were not touched.");
     }
     lines.push(`How voices are applied: ${sheet.treatment.summary}${sheet.treatment.speedChange ? "" : " No speed change."}`);
-    const rendersOwnVoice = sheet.treatment.kind === "render-owns-voice";
-    if (!rendersOwnVoice || sheet.treatment.usesElevenLabs === undefined) {
+    const playsPinnedVoices = sheet.treatment.kind !== "render-owns-voice";
+    const convertsClips = playsPinnedVoices && sheet.treatment.usage.clips > 0;
+    if (convertsClips || sheet.treatment.usesElevenLabs === undefined) {
         lines.push(`Cost: billed to your own ElevenLabs key. ${sheet.treatment.costNote} ` +
             `About ${sheet.treatment.usage.seconds} seconds of speech across ` +
             `${sheet.treatment.usage.clips} clips will be converted.`);

package/dist/commands/workflow.d.ts

@@ -386,6 +386,9 @@ export interface WorkflowRun {
     _id: string;
     workflowId: string;
     workflowName: string;
+    runTitle?: string;
+    title?: string;
+    moduleKey?: string;
     status: WorkflowRunStatus;
     error?: string;
     counts?: WorkflowCounts;
@@ -408,7 +411,12 @@ export interface WorkflowRun {
         approvedAt: number;
     }[];
     provenance?: RunProvenance;
+    pauseAhead?: {
+        nodeId: string;
+        after: "frames" | "storyboard";
+    };
 }
+export declare function pauseAheadLine(ahead: NonNullable<WorkflowRun["pauseAhead"]>): string;
 export type WorkflowRunProjection = Omit<WorkflowRun, "nodes"> & {
     nodes?: never;
 };

package/dist/commands/workflow.js

@@ -6,6 +6,7 @@ import { formatApiError } from "../lib/format.js";
 import { pollUntilDone } from "../lib/poll.js";
 import { normalizeRunStatus, isTerminalRunStatus, storedWorkflowStatusForms, TERMINAL_RUN_STATUSES, } from "../lib/runStatus.js";
 import { runVerdict } from "../lib/runVerdict.js";
+import { runDisplayName, runHeadingWord, runTypeWord } from "../lib/runNames.js";
 import { workflowToYaml, parseWorkflowText } from "../lib/workflowText.js";
 import { missingRouteLine } from "../lib/route-support.js";
 import { getChannel } from "../lib/channel.js";
@@ -366,6 +367,11 @@ export function formatRulesLine(provenance) {
         : "";
     return `Format rules: ${rigName}${version}${where}.${borrowed}`;
 }
+export function pauseAheadLine(ahead) {
+    return ahead.after === "frames"
+        ? "This run will pause for your approval once the frames are ready. The reference, cast and scene pictures are made first (a few cents each). Nothing bigger is spent until you approve."
+        : "This run will pause for your approval once the storyboard is written. Nothing is spent on pictures, voices or clips until you approve.";
+}
 const LIST_PATH = "/api/v2/workflows";
 const RUN_PATH = "/api/v2/workflows/run";
 const STATUS_PATH = "/api/v2/workflow";
@@ -1114,7 +1120,13 @@ const RECENT_RUNS_PAGE = 25;
 export function formatRecentRuns(runs) {
     if (runs.length === 0)
         return "No workflow runs found for the active brand.";
-    const rows = table(["workflow", "status", "created", "id"], runs.map((r) => [r.workflowName, runVerdict(r), dateOnly(r.createdAt), r._id]));
+    const rows = table(["run", "type", "status", "created", "id"], runs.map((r) => [
+        runDisplayName(r),
+        runTypeWord(r),
+        runVerdict(r),
+        dateOnly(r.createdAt),
+        r._id,
+    ]));
     const notes = [];
     if (runs.length >= RECENT_RUNS_PAGE) {
         notes.push(`Showing the ${RECENT_RUNS_PAGE} newest runs — there may be older ones this list doesn't reach.`);
@@ -1195,7 +1207,7 @@ export function formatImportSummary(result, mode = {}) {
 export function formatWorkflowRun(run) {
     const lines = [];
     const counts = formatCounts(run.counts);
-    lines.push(`Workflow run — ${run.workflowName}`);
+    lines.push(`${runHeadingWord(run)} — ${runDisplayName(run)}`);
     lines.push(`runId:        ${run._id}`);
     lines.push(`workflowId:   ${run.workflowId}`);
     if (run.triggerRunId)
@@ -1218,6 +1230,10 @@ export function formatWorkflowRun(run) {
     if (run.autoApprovals && run.autoApprovals.length > 0) {
         lines.push(`auto-approved: ${run.autoApprovals.length} checkpoint stop${run.autoApprovals.length === 1 ? "" : "s"} (${run.autoApprovals.map((a) => a.nodeId).join(", ")}) — launched with --auto-approve, nobody reviewed these`);
     }
+    if (run.pauseAhead) {
+        lines.push("");
+        lines.push(pauseAheadLine(run.pauseAhead));
+    }
     const parkedNodeId = gateParkedNodeId(run.status, run.pauseReason, run.pausedNodeId);
     if (run.nodes.length > 0) {
         lines.push("");

package/dist/lib/runNames.d.ts

@@ -0,0 +1,10 @@
+export declare const COPY_FACE_MODULE_KEY = "copy-face";
+export interface RunNameFields {
+    workflowName?: string;
+    runTitle?: string;
+    title?: string;
+    moduleKey?: string;
+}
+export declare function runDisplayName(run: RunNameFields): string;
+export declare function runTypeWord(run: RunNameFields): string;
+export declare function runHeadingWord(run: RunNameFields): string;

package/dist/lib/runNames.js

@@ -0,0 +1,10 @@
+export const COPY_FACE_MODULE_KEY = "copy-face";
+export function runDisplayName(run) {
+    return run.title || run.runTitle || run.workflowName || "";
+}
+export function runTypeWord(run) {
+    return run.moduleKey === COPY_FACE_MODULE_KEY ? "Copy run" : "Workflow";
+}
+export function runHeadingWord(run) {
+    return run.moduleKey === COPY_FACE_MODULE_KEY ? "Copy run" : "Workflow run";
+}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@aicopycoders/exodus",
-  "version": "2026.9.1900-beta.2",
+  "version": "2026.9.2000-beta.1",
   "description": "CLI + Claude Code skills for the Copycoders Exodus ad pipeline",
   "repository": {
     "type": "git",