@gentbajko/slopify 3.25.9 → 4.1.0

package/dist/tutorials/Reviews-and-Checkpoints.md

@@ -1,156 +1,164 @@
-# Reviews and Checkpoints
+# Pauses and Checks
 
-Two ways to keep an eye on a run before it spends more. A **review checkpoint** holds the run before a step until you approve it, so you can check and edit what came before. An **automatic review** has a reviewer model check each finished item (the article, every image, the narration, the thumbnail, every short) and flag it, or make it again, when it fails.
+Two ways to keep an eye on a run before it spends more. A **pause** holds the run after a step until you approve what it made, so you can read, listen and edit before the next paid step. An **AI check** has an AI look at each finished item (the article, every image, the narration, the thumbnail, every short) and flag it, or make it again, when it fails.
 
-**Where to find it:** Play → **Reviews** row → **Change**, which opens the **Review** side panel. **Review the whole setup** in the right rail and `Ctrl+Enter` open the same panel. After the run starts: the project page → **Checkpoints**, and Edit project → **Reviews**.
+**Where to find it:** Create → **Pauses and checks** row → **Change**. After the run starts: the project → **Edit settings** → **Pauses and checks**. A run held at a pause waits under Home → **Needs you**.
 
-## The Review side panel
+## The row on Create
 
-The panel holds, from top to bottom:
+The row's line says what is set, for example "Pauses for your approval after the article (before the narration); an AI checks 3 steps.", or "Runs straight through: no pauses for your approval, no AI checks." Opened, it holds **Pause for my approval** at the top and **Have an AI check the results** under it.
 
-1. Any setup errors. Press one to jump to the field.
-2. **Review checkpoints**.
-3. **Automatic reviews**.
-4. The whole setup in words, with **Read the resolved prompt** (or **Read the supplied article and resolved prompts**), which shows each prompt exactly as it will be sent.
+**See everything this run will do** in Create's rail lists the pauses with the rest of the run.
 
-The estimate and the Play key stay in the right rail. Opening the panel also asks the server to check the setup, since it knows rules the browser cannot see. **Refresh review** in the right rail does the same and recalculates the estimate now, instead of waiting for you to pause typing. Use it after changing a prompt in the Library.
+## Pauses
 
-The Reviews row summary says what is set, for example `1 checkpoint · 2 automatic reviews`, or `No checkpoints · no automatic reviews`.
+A pause holds the run after a step so you can check and edit what it made. Steps that do not depend on it carry on. You approve each pause to go on.
 
-## Review checkpoints
-
-A checkpoint holds the run before a step so you can check and edit what came before it. Steps that do not depend on it carry on. You approve each checkpoint on the project page to continue.
-
-| Checkpoint | Holds | Available when |
+| Pause | Holds | Offered when |
 | --- | --- | --- |
-| **Before Audio** | The narration and what depends on it | Narration Source is **Generate** |
-| **Before Images** | The images and what depends on them | Images Source is **Generate** |
-| **Before Video / export** | The render and the exports | The video is on, or narration is on |
+| **Pause after the article (before the narration)** | The narration and what depends on it | Narration is **AI voice** |
+| **Pause after the narration (before the images)** | The images and what depends on them (it reads "after the article" when there is no narration) | Images are **Made by AI** |
+| **Pause after the images (before the video)** | The video and the other files put together from it (it reads "before the sound file" without a video) | The video is on, or there is narration |
 
-A step must be generated to have a checkpoint. A greyed checkbox says why: "Choose Generate for Images to add this checkpoint", or "Enable Video or Audio to review before export." A ticked checkpoint whose step you later turn off says "This step is unavailable. Remove its checkpoint or enable the step."
+A pause is offered only for a step the run makes itself. A ticked pause whose step you later turn off stays, with the reason in red: "This run makes no images to pause before: untick it."
 
-Default: no checkpoints.
+Default: no pauses.
 
-### Add a checkpoint on Play
+### Add a pause on Create
 
-1. Open the **Reviews** row (or press **Review the whole setup**).
-2. Under **Review checkpoints**, tick **Before Audio**, **Before Images** or **Before Video / export**.
-3. Press **Refresh review** to see, per video, which steps each checkpoint also holds ("Run 1 · Before Images", "Also holds: Video / export.").
-4. Start the run.
+1. Open the **Pauses and checks** row.
+2. Under **Pause for my approval**, tick the pause you want.
+3. Press **Make the video**.
 
-Starting the run does not approve the checkpoints. The run stops at each one until you approve it on the project page.
+Starting the run does not approve the pauses. The run stops at each one until you approve it.
 
-### Continue a run held at a checkpoint
+### Go on after a pause
 
-When the run reaches a checkpoint, the held step waits and the project's next action says so.
+When the run reaches a pause, the held step waits, the project reads **Waiting for your review**, and it shows under Home → **Needs you**.
 
-1. Open the project. The right rail's next action names what runs after approval, for example **Approve and make the images**, **Approve and record the narration** or **Approve and render the video**.
+1. Open the project. The title row's next action names what runs after approval: **Approve and make the narration**, **Approve and make the images** or **Approve and make the video**.
 2. Check and edit what came before: read the article, listen to the narration, look at the images. See [Editing a Project](Editing-a-Project).
-3. Press the next action, or open **Checkpoints** in the project's side list (it shows how many are held) and press **Approve Images checkpoint** (or the stage you are at).
+3. Press the next action. The same button sits on the pause's card in **Edit settings** → **Pauses and checks**, and Home's **Needs you** has one too.
 
-The checkpoint then says "Checkpoint approved. Work will run when its dependencies are ready." Each checkpoint card shows its state:
+Each pause's card, **Paused after the article** (or the narration, or the images), says where it stands, with the steps that wait for it:
 
-| Message | Meaning |
+| Line | Meaning |
 | --- | --- |
-| "Images is held for your review. Its dependent work waits for this approval; independent work can continue." | Waiting for you. The project reads **Waiting for your review**. |
-| "Inputs changed. Review the current revision before approving." | You edited something the held step uses. Check the new version before approving. |
-| "Authorized for this revision. Work can run when dependencies are ready and the project is resumed." | Approved; it runs when it can. |
-| "Checkpoint satisfied. Its reviewed work is complete." | Done. |
+| "The article is ready for your review. The narration and what follows wait for your approval; other work goes on." | Waiting for you. |
+| "What it was made from has changed. Look at the current version before approving." | You edited something the held step uses. Check the new version before approving. |
+| "Approved for this version. What comes after it runs once what it needs is ready and the run goes on." | Approved; it runs when it can. |
+| "Approved, and what came after it is made." | Done. |
+| "Not reached yet." | The run has not got there. |
 
-If the project is paused, the panel says "The project is paused. Continue the run before approving held work." Pausing always applies on top of checkpoints.
+If the project is paused, the card says "The project is paused. Continue the run before approving held work." Pausing the project always applies on top of its pauses.
 
-A checkpoint is approved once. If it was already approved (by you in another tab, or by an [agent](Agents-MCP)), approving again changes nothing and says so.
+A pause is approved once. If it was already approved (by you in another tab, or by an [agent](Agents-MCP)), approving again changes nothing and says "It was already approved."
 
-### Change checkpoints after the run started
+### Change the pauses after the run started
 
-On the project page → **Checkpoints** → **Checkpoint choices**:
+In the project → **Edit settings** → **Pauses and checks** → **Pauses for this run**:
 
-1. Tick a step that has not started to add a checkpoint, or untick one to remove it. Work already admitted for an unticked step carries on when ready.
-2. Press **Save checkpoints**.
+1. Tick a step that has not started to add a pause, or untick one to remove it. Work already let through for an unticked step carries on when ready.
+2. Press **Save pauses** (it appears once you change a tick). They take effect at once, without **Save changes**.
 
-Only generated steps can have one. If the project was edited elsewhere in the meantime, press **Reload checkpoint choices** and make your change again.
+Only steps the run makes itself, and has not started, can be ticked. Once the run has finished there is nothing left to pause: the ticks give way to the line "Nothing is left to pause: every step a pause can hold has run." Making something again starts once you have seen what it costs, without a pause. If the project was edited elsewhere in the meantime, press **Load the pauses again** and make your change again.
 
-## Automatic reviews
+## AI checks
 
-A reviewer checks each finished item and saves its verdict and reasons on the project page. The reviewer is a model Slopify asks (each review is a text-model call, counted in usage and in the estimate), or an agent you connected (free; see [The External agent reviewer](#the-external-agent-reviewer)).
+An AI checks each finished item and saves its verdict and reasons on the project page. The checker is a model Slopify asks (each check is one AI call, counted in usage and in the estimate), or an agent you connected (free; see [Your own AI agent as the checker](#your-own-ai-agent-as-the-checker)).
 
-| Option | What it does | Default / range |
-| --- | --- | --- |
-| **Who reviews** | **Model**: Slopify asks the reviewer model below, one paid call per item. **External agent (MCP)**: an agent connected in Settings → Connections → Agents (MCP) answers each review instead, at no API cost. | Model |
-| **Reviewer** | Who runs the review model (greyed for **External agent (MCP)**). Picture reviews (images, the thumbnail and shorts) need Claude Code or Codex, which can look at images; other providers review text only. | None |
-| **Reviewer model** | The model that judges each item. A stronger model catches more but costs more per review. It can differ from the model that wrote the text. | None |
-| **Redos per item** | How many times **Flag and redo** makes a failed item again before it keeps the last one and flags it. Each redo costs the item again plus one more review. | 2; 0 to 5 |
+**Have an AI check the results** sets every stage the run makes at once:
 
-Then, for each stage (**Article**, **Images**, **Narration**, **Thumbnail**, **Shorts**), pick a mode and a prompt:
-
-| Mode | What happens to a failed item |
+| Choice | What happens to a failed item |
 | --- | --- |
-| **Off** (default) | No review. |
-| **Flag only** | The item is marked failed, with the reasons, and the run carries on. |
-| **Flag and redo** | The item is made again and reviewed again, up to **Redos per item**. After that it is kept and flagged. It never loops. |
+| **Off** (default) | No check. |
+| **Flag problems** | The item is marked failed, with the reasons, and the run carries on. |
+| **Flag and redo** | The item is made again and checked again, up to **Redos per item**. After that it is kept and flagged. It never loops. |
+
+The line under it says "Each check is one more AI call per item; a redo makes the item again, at its price." Turning checks on with no checker picked uses the AI that writes the article. When the stages are set differently, the choice shows none of the three and the line says "Each stage is set on its own under Checker and stages."
+
+### Checker and stages
 
-The prompt next to each mode is **Built-in** or a **Review** prompt from the Library (see [Prompts](Prompts)). It is greyed while the mode is Off. A stage the run makes nothing for is greyed too; a stage set to review but with nothing to review says "This stage makes nothing to review in this run."
+**Checker and stages** under it folds out the rest. Its summary says who checks and how many stages are checked.
 
-### What each review checks
+| Option | What it does | Default / range |
+| --- | --- | --- |
+| **Who checks** | **An AI model**: Slopify asks the model under **Checked by**, one paid call per item. **Your own AI agent**: an agent connected in Settings → **AI agents** answers each check instead, at no cost. | An AI model |
+| **Redos per item** | How many times **Flag and redo** makes a failed item again before it keeps the last one and flags it. Each redo costs the item again plus one more check. | 2; 0 to 5 |
+| **Checked by** | The **Service** and **Model** that judge each item. A stronger model catches more but costs more per check. It can differ from the one that wrote the text. Checks of the images, the thumbnail and shorts need Claude Code or Codex, which can look at images; other services check text only. | The article's writer |
+
+Then the table **Stage** / **Check, and the prompt it uses** sets each stage, in the order the run makes them (**Article**, **Promise check (before narration)**, **Narration**, **Images**, **Thumbnail**, **Shorts**), on its own: **Off**, **Flag problems** or **Flag and redo**, and, once it is on, the prompt: **Built-in prompt** or one of your **Review** prompts (see [Prompts](Prompts)). A stage the run makes nothing for is left out; the promise check stays, greyed, saying it "Needs an article the AI writes."
+
+### What each check looks at
 
-| Stage | The built-in reviewer checks |
+| Stage | The built-in check looks at |
 | --- | --- |
 | **Article** | The article prompt's rules (length, structure, tone), no invented lore, no repeated sections. The research notes are included. |
+| **Promise check (before narration)** | That the opening keeps what the title and thumbnail promise. See [The promise check](#the-promise-check). |
 | **Images** | Malformed hands or bodies, stray text or letters, that it matches its brief, and that it matches the establishing image when one is on. |
-| **Narration** | Nothing missing or garbled compared with the text: the reviewer reads what the word timing heard and the stretches it could not hear. |
+| **Narration** | Nothing missing or garbled compared with the text: the checker reads what the word timing heard and the stretches it could not hear. |
 | **Thumbnail** | Readable at phone size, clear subject. |
 | **Shorts** | The short's images at phone size: clear subject, no stray text, fits the title. |
 
-### Set up automatic reviews
+### The promise check
+
+One cheap AI call before the paid narration. The checker reads the video's title, the thumbnail brief (when the thumbnail is drawn from a prompt), the opening as the voice will say it (about the first 30 seconds: 75 words at 150 words a minute) and the whole article.
+
+- **Passes** when the opening starts delivering what the title and thumbnail promise and gives a reason to keep watching.
+- **Fails** on a slow preamble, a greeting or channel housekeeping (subscribe, welcome back) before the subject, an opening about something else, or a promise the article never keeps. The reasons name the sentence that should come first.
+- **Flag and redo** writes the article again and checks the new one. The narration, the pictures and everything else that reads the article wait for the answer, so nothing paid is made from an opening that failed.
+- It needs an article the AI writes: otherwise its row is greyed and says "Needs an article the AI writes."
+- It uses the checker above and adds one AI call to the estimate. **Have an AI check the results** leaves it as it is; set it on its own under Checker and stages. Default: Off.
+
+### Set up the AI checks
 
-1. Open the **Reviews** row.
-2. Pick the **Reviewer** and **Reviewer model**. For image, thumbnail or shorts reviews, pick Claude Code or Codex. With another reviewer, the panel says "Picture reviews need Claude Code or Codex as the reviewer: Slopify can't show pictures to this one." and the run cannot start.
-3. Optional: set **Redos per item** (empty means 2).
-4. For each stage you want checked, pick **Flag only** or **Flag and redo**, and optionally a Review prompt.
-5. Check the estimate: each review adds one text-model call. Redos are not in the estimate; each costs what the item cost, at most the redos allowed.
+1. Open the **Pauses and checks** row.
+2. Under **Have an AI check the results**, pick **Flag problems** or **Flag and redo**.
+3. Optional: open **Checker and stages** to pick who checks (**Checked by**, or **Your own AI agent**), set **Redos per item** (empty means 2), or set a stage on its own with a Review prompt. For image, thumbnail or shorts checks, pick Claude Code or Codex. With another checker, the row says "Checks of images need Claude Code or Codex as the checker: Slopify can't show images to this one." and the run cannot start.
+4. Check the estimate: each check adds one AI call. Redos are not in the estimate; each costs what the item cost, at most the redos allowed.
 
-### The External agent reviewer
+### Your own AI agent as the checker
 
-With **Who reviews** set to **External agent (MCP)**, no model is called: each review waits for an agent connected to Slopify to answer it (see [Agents (MCP)](Agents-MCP)). Every stage can use it, picture stages included, since the agent is shown the pictures. It adds nothing to the estimate.
+With **Who checks** set to **Your own AI agent**, no model is called: each check waits for an agent connected to Slopify to answer it (see [Agents (MCP)](Agents-MCP)). Every stage can use it, picture stages included, since the agent is shown the pictures. It adds nothing to the estimate.
 
-- While a review waits, what depends on the item waits, as with any review. If nothing else holds the run, the project reads **Waiting for the agent** (header, Projects, Home's Running now, the calendar). It is not in **Needs you**, sends no "Waiting for you" notice and has no **Continue the run**: it goes on by itself.
-- The agent answers **pass** or **fail**. A fail needs at least one reason, at most 10, each up to 500 characters. The answer then follows the stage's mode like a model's verdict: **Flag only** flags it, **Flag and redo** makes it again up to **Redos per item**.
-- The first answer saved wins. A second answer to the same review is refused ("This review was already answered").
-- **No answer in 2 hours:** the review passes by itself. Its badge reads **Passed without review**, the work goes on and a notification says "Passed without review: no agent answered in 2 h". The deadline is saved, so it survives a restart. A paused project's review still passes at its deadline, but nothing runs until you continue the run.
-- **Canceled runs:** canceling stops the run's waiting reviews, and they never pass. Taking the step up again asks again, with a fresh two hours.
+- While a check waits, what depends on the item waits, as with any check. If nothing else holds the run, the project reads **Waiting for the agent** (title row, Projects, Home's Running, the calendar). It is not in **Needs you**, sends no "Waiting for you" notice and has no **Continue the run**: it goes on by itself.
+- The agent answers **pass** or **fail**. A fail needs at least one reason, at most 10, each up to 500 characters. The answer then follows the stage's choice like a model's verdict: **Flag problems** flags it, **Flag and redo** makes it again up to **Redos per item**.
+- The first answer saved wins. A second answer to the same check is refused ("This review was already answered").
+- **No answer in 2 hours:** the check passes by itself. Its badge reads **Passed without review**, the work goes on and a notification says "Passed without review: no agent answered in 2 h". The deadline is saved, so it survives a restart. A paused project's check still passes at its deadline, but nothing runs until you continue the run.
+- **Canceled runs:** canceling stops the run's waiting checks, and they never pass. Taking the step up again asks again, with a fresh two hours.
 
-### Reviews on the project page
+### Verdicts on the project page
 
-Beside the article, the narration, the thumbnail and each short, and on each image tile, the verdict shows as **Review passed**, **Flagged by review**, **Waiting to be made again**, **Being made again**, **Redone after review** or **Accepted by you**, with the reasons. With the External agent reviewer, or when an agent acted on it, it reads **Passed by agent**, **Flagged by agent**, **Redone after agent review**, **Accepted by agent**, **Made again by agent**, or **Passed without review** when no agent answered in time. For a failed item nobody has acted on:
+Beside the article (with the promise check's verdict under the article's), the narration, the thumbnail and each short, and on each image, the verdict shows as **Review passed**, **Flagged by review**, **Waiting to be made again**, **Being made again**, **Redone after review** or **Accepted by you**, with the reasons. With your own AI agent as the checker, or when an agent acted on it, it reads **Passed by agent**, **Flagged by agent**, **Redone after agent review**, **Accepted by agent**, **Made again by agent**, or **Passed without review** when no agent answered in time. For a failed item nobody has acted on:
 
 - **Overrule** accepts it as it is.
-- **Redo** makes it again (a new version of the project, then the item and what depends on it are rebuilt), and it is reviewed again.
+- **Redo** makes it again (a new version of the project, then the item and what depends on it are made again), and it is checked again.
 
 A flagged item takes one decision. If it was already decided (by an agent, for example), pressing **Overrule** or **Redo** changes nothing and says how it was decided.
 
-Both buttons wait while other work on the project is running, with one exception: while a **Flag and redo** review's automatic redo is still waiting to start (**Waiting to be made again**), **Overrule** stays available and calls the redo off. The item is accepted as it is, the badge turns to **Accepted by you**, and the work waiting on it goes ahead. Once the redo has started, Overrule is greyed: "Slopify is already making this item again; wait for the new version".
+Both buttons wait while other work on the project is running, with one exception: while a **Flag and redo** check's redo is still waiting to start (**Waiting to be made again**), **Overrule** stays available and calls the redo off. The item is accepted as it is, the badge turns to **Accepted by you**, and the work waiting on it goes ahead. Once the redo has started, Overrule is greyed: "Slopify is already making this item again; wait for the new version".
 
-When an automatic review flags an item and keeps it (its redos ran out, or the stage is on **Flag only**), Slopify sends a **Review needs a decision** notification (see [Notifications](Notifications)), and the project shows in the Calendar's **Needs you** with **Open to review** (see [Calendar](Calendar#needs-you)). When an agent's review flagged it, the notification ends in **· by agent**. Everything that uses a reviewed item waits for its review, in the same way a checkpoint holds work. Reviewing the same output with the same reviewer and prompt again reuses the earlier verdict.
+When a check flags an item and keeps it (its redos ran out, or the stage is on **Flag problems**), Slopify sends a **Review needs a decision** notification (see [Notifications](Notifications)), and the project shows under Home → **Needs you**. When an agent's check flagged it, the notification ends in **· by agent**. Everything that uses a checked item waits for its check, in the same way a pause holds work. Checking the same file with the same checker and prompt again reuses the earlier verdict.
 
-## Checkpoints or reviews?
+## Pauses or checks?
 
 | You want to | Use |
 | --- | --- |
-| Read the article yourself before paying for the narration | **Before Audio** checkpoint |
-| Look at the images before the render | **Before Images** is too early; use **Before Video / export** |
-| Catch broken hands or stray text without watching every run | **Images** review, **Flag and redo** |
-| Run unattended on a schedule and fix things later | Reviews on **Flag only**, no checkpoints |
-| Have reviews done without paying a reviewer model | **External agent (MCP)** with an hourly agent routine; see [Agents (MCP)](Agents-MCP) |
+| Read the article yourself before paying for the narration | **Pause after the article (before the narration)** |
+| Look at the images before the video is put together | **Pause after the images (before the video)** |
+| Catch broken hands or stray text without watching every run | **Have an AI check the results**: **Flag and redo** |
+| Run unattended on a schedule and fix things later | Checks on **Flag problems**, no pauses |
+| Have the checks done without paying for a model | **Your own AI agent** with an hourly agent routine; see [Agents (MCP)](Agents-MCP) |
 
 ## Tips
 
-- Checkpoints and reviews are saved with a template, so every video started from it gets them. See [Templates](Templates).
-- Keep **Redos per item** low for images: five redos is already six paid images for one slot.
-- You can change reviews later in Edit project → **Reviews**. See [Editing a Project](Editing-a-Project).
+- Pauses and checks are saved with a template, so every video started from it gets them. See [Templates](Templates).
+- Keep **Redos per item** low for images: five redos is already six paid images for one picture.
+- You can change the checks later in the project's **Edit settings** → **Pauses and checks** (then **Save changes**). See [Editing a Project](Editing-a-Project).
 
 ## Related pages
 
-- [Play Overview](Play-Overview)
+- [Create Overview](Create-Overview)
 - [Project Page](Project-Page)
 - [Editing a Project](Editing-a-Project)
 - [Recovery and Retries](Recovery-and-Retries)

package/dist/tutorials/Schedules.md

@@ -1,6 +1,6 @@
 # Schedules
 
-A schedule starts a new project from a saved template at a set local time: every day, on chosen weekdays, or once. Give it a list of topics and each run takes the next one. Turn on topic generation and it asks your text model for new topics before the list runs dry.
+A schedule starts a new project from a saved template at a set local time: every day, on chosen weekdays, or once. Give it a list of topics and each run takes the next one. Turn on topic generation and it asks your AI for new topics before the list runs dry.
 
 **Where to find it:** **Calendar → Schedules**, the second tab on the calendar. Old `/schedules` links and **G then S** open it too. A channel's **Schedules** tab lists the schedules that run its templates.
 
@@ -15,7 +15,7 @@ A schedule starts a new project from a saved template at a set local time: every
 1. Open **Calendar**, then the **Schedules** tab, and press **New schedule**.
 2. Fill in **Name** (up to 200 characters). It never appears in the videos.
 3. Pick the **Template** each run starts from. The schedule keeps the template's current version.
-4. Choose the **Cadence** and times (see below).
+4. Choose **How often** and the times (see below).
 5. Add **Topics** (see below), or leave the list empty to run the template exactly as saved.
 6. Optional: set up **Topic generation**, a **Spend ceiling** and what happens to a **Missed run**.
 7. Press **Save schedule**.
@@ -26,7 +26,7 @@ The schedule appears under **Saved schedules** with its status, next run and the
 
 | Setting | What it does | Default |
 |---|---|---|
-| **Cadence** | **Every day** runs once a day at the local time. **Selected weekdays** runs on the days you tick. **One time** runs once at a date and time, then shows as completed. Each run starts one project. | Every day |
+| **How often** | **Every day** runs once a day at the local time. **Selected weekdays** runs on the days you tick. **One time** runs once at a date and time, then shows as completed. Each run starts one project. | Every day |
 | **Local time** | The clock time each run starts, read in the timezone below. | 09:00 |
 | **Weekdays** | The days a Selected weekdays schedule runs. Tick at least one; each ticked day is one project a week. | Mon, Wed and Fri |
 | **Run at** | The date and time of a One time schedule's single run. | |
@@ -44,7 +44,7 @@ Each run takes the first topic, starts one project with it and removes it from t
 
 **Each topic fills** is the template keyword a topic replaces, such as `{{Topic}}`. By default it is the keyword the template's project title uses, or else the first one. If the title doesn't use this keyword, every project gets the same title.
 
-Every other keyword of the template has an every-run value: the value it gets on every run, such as a word count. The fields are labelled `<keyword> (every run)` and use the same keyword list as Play, Edit project and templates: the line under each says which prompts it feeds ("Feeds …"), or "Not used by any picked prompt". The default is the template's own value. A topic can set its own value for any keyword, which wins for that one run; the field then notes "Unless a topic sets its own."
+Every other keyword of the template has an every-run value: the value it gets on every run, such as a word count. The fields are labelled `<keyword> (every run)` and use the same keyword list as Create, a project's Edit settings and Edit template: the line under each says which prompts it feeds ("Feeds …"), or "Not used by any picked prompt". The default is the template's own value. A topic can set its own value for any keyword, which wins for that one run; the field then notes "Unless a topic sets its own."
 
 ### Three ways to write the list
 
@@ -83,39 +83,39 @@ Pick a schedule in the list to see its **Queued topics · N**. You can change th
 
 Each change saves at once and says so, for example "Renamed to “Hypatia of Alexandria”.", with **Undo** on the notice. Only the queue changes; the next run and the other settings stay as they are. If the schedule changed in another tab meanwhile, the change is refused with the reason; reload and try again.
 
-## Prepare a video ahead
+## Make a video ahead
 
-A scheduled video can be made ahead of its day, so the day only renders it: useful before a trip, or to read the article and listen to the narration first.
+A scheduled video can be made ahead of its day, so the day only puts the video together: useful before a trip, or to read the article and listen to the narration first.
 
 1. Open **Calendar** and find the topic's run (either view).
-2. Press **Prepare**.
+2. Press **Make ahead**, read what it does and spends, and press **Make ahead (paid)**.
 
-Slopify makes the project now, exactly as the schedule would on the day (the template as it is now, the topic and the schedule's values), and runs everything but the video: research, the article, narration, images, the thumbnail and the PDF. It then holds at a **Before Video** checkpoint, and the run shows **Prepared** with **Open**.
+Slopify makes the project now, exactly as the schedule would on the day (the template as it is now, the topic and the schedule's values), and runs everything but the video: research, the article, narration, images, the thumbnail and the PDF. It then pauses before the video, and the run shows **Waiting for its day** with **Open**.
 
 On the scheduled day the schedule finds that project by its title and continues it instead of making another: the hold is lifted and the video, its captions, shorts and YouTube description are made. The same goes for an unfinished project you made yourself under the same title. A finished one counts as a video already made, and the day makes a new one.
 
-- You can open the prepared project and change anything before its day, for example with Edit project.
-- A **Before Video** checkpoint the template already has stays: the day waits for you to approve it, and the project reads **Waiting for your review** and is listed under **Needs you** (and for a connected [agent](Agents-MCP)). A project held only by Prepare's own hold reads **Prepared** and is not listed as waiting anywhere: it waits for its day, not for you.
-- A setup that makes no video has nothing to hold back, so Prepare runs it to the end.
-- A schedule without topics gives every run the same title, so only a prepared project continues there.
-- The spend limit applies to Prepare as it does to a scheduled run.
+- You can open the project made ahead and change anything before its day, for example with Edit settings.
+- A pause before the video that the template already has (**Pause after the images (before the video)**) stays: the day waits for you to approve it, and the project reads **Waiting for your review** and is listed under **Needs you** (and for a connected [agent](Agents-MCP)). A project held only by Make ahead's own hold reads **Waiting for its day** and is not listed as waiting anywhere: it waits for its day, not for you.
+- A setup that makes no video has nothing to hold back, so Make ahead runs it to the end.
+- A schedule without topics gives every run the same title, so only a project made ahead continues there.
+- The spend limit applies to Make ahead as it does to a scheduled run.
 
 ## Topics that find themselves
 
-Under **Topic generation**, a schedule can ask an LLM for its next topics so it never runs dry.
+Under **Topic generation**, a schedule can ask the AI for its next topics so it never runs dry.
 
 | Setting | What it does | Default |
 |---|---|---|
 | **New topics** | **Off: add topics yourself.** **Generate and queue directly:** new topics join the end of the list. **Generate and hold for approval:** they wait under **Topics waiting**, and beside the calendar, until you approve them. | Off |
-| **Series brief** | What the series covers, its style and what makes a topic worth a video. Only topic generation reads it. Leave it empty to use the channel's series brief from the channel's **Brand** tab. Up to 4,000 characters. | Empty |
+| **Series brief** | What the series covers, its style and what makes a topic worth a video. Only topic generation reads it. Leave it empty to use the channel's series brief from the channel's **Name and look** tab. Up to 4,000 characters. | Empty |
 | **Keep at least this many queued** | Slopify asks for more topics whenever the queue, plus topics waiting for approval, drops below this number. From 1 to 100. | 10 |
-| **Topic generation LLM** | Ticked, it uses the provider and model the template writes with. Untick to pick another **Provider** and **Model**, such as a cheaper one. | The template's LLM |
+| **Use the template's AI model** | Ticked, it uses the service and model the template writes with. Untick to pick another **Service** and **Model**, such as a cheaper one. | Ticked |
 
 ### How generation works
 
 1. When the queue (plus held topics) drops below your number, Slopify starts one generation in the background. It checks about every 15 seconds.
-2. It makes one LLM call for what is missing plus 5 spare, at most 50 topics.
-3. The call carries the series brief and every title already known: this schedule's queued, held, rejected and used topics, the titles of the projects in the schedule's own channel, and that channel's **Existing videos** (see [Channels](Channels)). Another channel's videos don't count: two channels may cover the same subject.
+2. It makes one AI call for what is missing plus 5 spare, at most 50 topics.
+3. The call carries the series brief and every title already known: this schedule's queued, held, rejected and used topics, the titles of the projects in the schedule's own channel, and that channel's **Titles to avoid** (see [Channels](Channels)). Another channel's videos don't count: two channels may cover the same subject.
 4. Suggestions come back ranked most view-worthy first. Slopify drops near-duplicates of titles it already knows, for example "Hypatia" and "Hypatia's Library" count as one video, while "Ptolemaic Egypt" and "Roman Egypt" are two. A topic of one or two words only counts as made when a part of a title (split at `:`, `|`, brackets, `?` and similar) holds exactly those words, leaving aside words like "who", "what" or "explained": "Cleopatra" matches "Who was Cleopatra? The Last Pharaoh", but "Egypt" does not match "The Gods of Ancient Egypt".
 5. The rest join the queue, or wait for you under **Topics waiting**.
 
@@ -123,7 +123,25 @@ Only one generation per schedule runs at a time. A failed try shows its reason o
 
 ### Generate topics now
 
-Press **Generate topics now** to ask for new topics at once: at least 5, even when the queue is full and even during the 5-minute wait after a failed try. It is one LLM call. Saving the schedule also clears the wait.
+Press **Generate topics now** to ask for new topics at once: at least 5, even when the queue is full and even during the 5-minute wait after a failed try. It is one AI call. Saving the schedule also clears the wait.
+
+### Rank topics
+
+Under Edit → Topic generation, **Rank topics** orders the topics Slopify generates, best first, and shows a **Score** (0 to 100) on each queued and held topic, with why it scored so behind the ⓘ beside it:
+
+| Choice | What it reads | What leaves this computer |
+| --- | --- | --- |
+| **Off** (default) | Nothing: no scores, and topics join in the order they arrive, as before | Nothing |
+| **By what the library already covers** | How far the topic is from everything made, queued, held or turned down in this channel ("Library: new; nearest “…” (40% overlap)") | Nothing |
+| **Also by public search interest** | The library, plus how many of Google's search suggestions name the topic (DuckDuckGo's when Google doesn't answer) and its Wikipedia article's views last month | Each new topic's title, sent to Google, DuckDuckGo and Wikipedia |
+
+- When a generation finds more topics than it needs, the best-scored are kept. They join after the topics you added or moved, in score order; held topics are listed by score.
+- A topic you move keeps its place from then on.
+- With no connection, topics are still added, ranked by the library alone, and their reasons say "Search interest unknown: no connection".
+- **Rescore**, beside Queued topics, scores every queued and held topic again now and puts them in order; a topic you added gets a score but stays where it is.
+- It works with **New topics** Off too: Rescore then scores the topics you added, which keep the places you gave them.
+- With Rank topics Off, no scores show.
+- It never uses your channel's analytics.
 
 ### Approve held topics
 
@@ -138,28 +156,29 @@ Waiting topics count toward **Keep at least**, so nothing more is asked for whil
 
 ## Manage a schedule
 
-Pick a schedule in the list to see its detail: **When**, **Timezone**, **Next run**, **Topics** left, **Missed runs**, **Overlap**, **Spend ceiling** and its **Run history**.
+Pick a schedule in the list to see its detail: **When** (a weekly one names its days, for example "Mon, Wed, Fri at 09:00"), **Timezone**, **Next run**, **Topics** left, **Template** (its name, and which version runs), **Missed runs**, **Overlap**, **Spend ceiling** and its **Run history**. With two or more channels, the list shows the channel picked under **Showing:** in the page header.
 
 | Action | What it does |
 |---|---|
-| **Edit** | Opens the **Edit schedule** form on the right. Only an active or paused schedule can change. Press **Save changes**. |
-| **Pause** / **Resume** | Stops runs until you resume. Only an active schedule can pause. |
+| **Edit** | Opens the **Edit schedule** form on the right. Offered on an active or paused schedule, the ones that can change. Press **Save changes**. |
+| **Pause** / **Resume** | Stops runs until you resume. **Pause** is offered on an active schedule. |
 | **Cancel schedule** | Stops future runs for good. Existing projects and run history are kept. |
-| **Delete schedule** | Moves the schedule to **Settings → Backup & storage → Trash** for 30 days. Its run history is kept. Restore brings it back paused. |
-| **Filter schedules by channel** | Shows only one channel's schedules. |
+| **Delete schedule** | Moves the schedule to **Settings → Trash** for 30 days. Its run history is kept. Restore brings it back paused. Offered once the schedule is completed or canceled. |
 | **Switch to <model>** | Shown under a schedule whose template picks a retired model, with "<model> is retired." Switches that model to its replacement in one press, as in Settings → **Models**. See [Models](Models). |
 
 **Run history** lists each run with its status (running, done, failed, skipped and so on), the time it was due, and either the project it made or the reason it didn't run. Deleted schedules are listed under **Deleted schedules** at the bottom; pick one to see its history.
 
 ## When you edit the template
 
-A schedule always uses its template as it is now: save a change to the template and the next run, the Calendar and topic generation use it. If the change adds a keyword the schedule has no value for, fill it in under the schedule's topics or every-run values.
+A schedule runs the newest version of its template, so a change you save applies from the next run, as before. The schedule's **Template** line names it and says **runs each version you save**, for example "History starter · runs each version you save".
+
+That changes once you approve a version for schedules in [Edit template](Template-Studio#approve-for-schedules) (**More → Approve for schedules**) for the first time. From then on the schedule runs the newest approved version (or one Slopify carried over from it), and a version you save later waits until you look it over and approve it. The line then says, for example, "History starter · runs version 6 · version 7 waits for your approval". The schedule is never stopped or switched to an unchecked version by itself. If the change adds a keyword the schedule has no value for, fill it in under the schedule's topics or every-run values.
 
 ## Tips
 
-- Try the template once from Play before scheduling it, so you know what one run costs. Then set the **Spend ceiling** a little above that.
+- Try the template once from Create before scheduling it, so you know what one run costs. Then set the **Spend ceiling** a little above that.
 - Hold generated topics for approval at first. Once you trust the suggestions, switch to queue directly.
-- Write a clear series brief: what the series covers, the tone, and what makes a topic worth a video. Add your older videos under the channel's **Existing videos** so they aren't suggested again.
+- Write a clear series brief: what the series covers, the tone, and what makes a topic worth a video. Add your older videos under the channel's **Titles to avoid** so they aren't suggested again.
 
 ## Related pages