Changing a slide after installation
This page answers one question: “I changed my slide. Do I need to copy my code again?” Most changes reach your course as soon as you save the slide. A few are frozen in the code you installed. Use the tables below before you edit a slide that learners already use.
How it works
Section titled “How it works”The code you install holds the slide’s config UID and a few choices made when you copied it, such as the activity type and the variable names. Everything else is read from SlideMind each time a learner sends a message or opens the widget. So:
- A change SlideMind reads at each call applies right away, in every course where the slide is installed, without republishing the course.
- A change frozen in the installed code applies only after you copy the code again (or download the folder again for the Storyline Web Object) and republish your course.
A ready-made Storyline slide holds the same code as Copy the code, in the trigger of its button. When a change needs a new copy, download the slide again and import it with Merge, or paste the new code from Manual setup into that trigger. The slide also shows the question of a Q&A, or the opening message of a Chat, as it was when you downloaded it: edit the text in Storyline, or download the slide again.
Changes that apply right away
Section titled “Changes that apply right away”For every destination
Section titled “For every destination”- The instructions, the question shown to the learner, the course and module context, and the response example.
- The opening message of a Chat, as the AI receives it.
- The AI model and its response settings: provider, temperature, max tokens and reasoning effort.
- What the AI does on the last turn of a Chat when you choose the AI wraps up.
- The exchange memory settings under Remember for later slides, such as Keep the learner’s answer or Keep an AI summary.
- The slide status. Setting a slide back to Draft or Archived stops the answers in your course at once.
- Plain-text AI responses in Settings › AI & models.
- Your allowed websites, and how the workspace pays for AI (SlideMind Credits or your own provider key).
For SlideMind widget destinations (Storyline Web Object, Rise 360, HTML widget)
Section titled “For SlideMind widget destinations (Storyline Web Object, Rise 360, HTML widget)”- Everything in the Look step: appearance, Q&A layout, labels, AI notice and Learner language.
- The opening message of a Chat, its turn limit and what happens after the last message.
- The slide title shown in the widget header, unless you typed a Widget title.
The widget address in newly generated code carries only the activity type, plus the scope for a capture. Older snippets that also carry appearance settings keep those settings: as the app says, “replace them to follow the slide.”
Changes frozen in the installed code
Section titled “Changes frozen in the installed code”| Change | Authoring tool code (Storyline JavaScript, Captivate, Lectora, dominKnow | ONE, iSpring Suite) | Storyline Web Object | Rise 360 and HTML widget |
|---|---|---|---|
| Activity type (Q&A, Chat, Capture) | Copy again | Download again (a Capture always uses the Storyline JavaScript code) | Copy again |
| Name of the question or answer variable, including a change of slide title or number when you never renamed the variables | Copy again | Download again (answer variable) | Not applicable |
| Add a value sent from the authoring tool (Values from Storyline, for example) | Copy again | Not available | Not available |
| Add or rename a return variable | Copy again. Until then, the new field reaches only the …_json variable, and a renamed field stops filling its old variable | Applies right away. Create the matching Storyline variable and republish your course | Not applicable. In a Q&A, the score and verdict settings in Look apply right away |
| The Text field the answer variable receives (Field sent to the reply variable, or the first Text field) | Copy again | Applies right away | Not applicable |
| Chat turn limit | Copy again | Applies right away | Applies right away |
| Capture slide: the values to remember (memory key, authoring tool variable or scope) | Copy again | Not applicable | Not applicable |
| Capture scope chosen in the Installation step | Copy again | Not applicable | Not applicable |
| Capture variables during the AI call: add a row, or change the authoring tool variable of a row | Copy again | Not applicable | Not applicable |
| Capture variables during the AI call: change the memory key or scope of an existing row | Applies right away | Not applicable | Not applicable |
| An update to the integration code itself (see below) | Copy again | Download again | Copy again |
Removing a value sent from the authoring tool does not need a new copy: the installed code keeps sending it and the AI ignores it unless your instructions still use it.
Other changes
Section titled “Other changes”- Changing the destination. Install the code of the new destination and remove the old one from your course. For a Chat, the destination also changes how SlideMind answers right away: see the fourth pitfall below.
- Duplicating a slide. The copy gets its own config UID and starts as a draft. Variable names you renamed on the original are not copied. Install the copy’s own code.
- Deleting a slide. Every installed code that uses its config UID stops responding.
- Restoring an earlier version from History. The restored settings apply right away, and the code already installed keeps working. If the restored version has a different activity type or variable names, the table above applies.
Updates that need a new copy
Section titled “Updates that need a new copy”SlideMind sometimes improves the integration code itself. Code you installed earlier keeps working, but it gets the improvement only once you copy it again (or download the folder again). Updates are announced in What’s new.
| Date | Destination | What a new copy brings |
|---|---|---|
| September 22, 2026 | Authoring tool code, Chat slides | The conversation history. Create the …_history and …_done variables, then copy the code again. Older code sends one message at a time. |
| September 25, 2026 | Authoring tool code | Quiet retries when the service is busy. |
| September 26, 2026 | Rise 360 | The completion signal used by Set completion requirements. |
| October 7, 2026 | Storyline Web Object, Chat slides | The structured fields of a Chat, written to Storyline variables after every reply. Older folders write only the reply and …_done. |
| October 17, 2026 | Authoring tool code and Storyline Web Object, with structured results | Each field is written to <answer>_<field>, for example aiResp_s2_quiz_score, instead of a variable named like the field, and the answer variable of the authoring tool code receives the feedback text instead of the raw JSON. Rename your return variables in your course when you install the new code. |
| October 17, 2026 | Storyline code | The …_busy variable for a loading layer, the error variable emptied before each send, no second send while the AI answers, and a time limit for a request that never answers. |
| October 26, 2026 | Authoring tool code and Storyline Web Object, Chat slides | The turn counter variables …_turns and …_max_turns. Create them only if you show a counter, then copy the code or download the folder again. |
Pitfalls to avoid
Section titled “Pitfalls to avoid”1. Generated names follow the slide title and number
Section titled “1. Generated names follow the slide title and number”Until you rename them, the question and answer variable names are built from the slide number and title, for example userMsg_s2_practiceRolePlayA and aiResp_s2_practiceRolePlayA.
- If you rename the slide or change its number, the code you already installed keeps working.
- But the next code you copy uses new names, which no longer match the variables you created in your authoring tool.
To avoid this, rename the variables in the Installation step before the first copy. Type a name of your own: a name identical to the generated one keeps following the title.
2. The turn limit is frozen in authoring tool code
Section titled “2. The turn limit is frozen in authoring tool code”In the authoring tool interface, the Chat turn limit is written into the code. If you change it in SlideMind, the installed code keeps the old limit until you copy it again. This is easy to miss, because everything else in a Chat applies right away.
Until you copy the code again, the two limits disagree. Say the code allows 5 turns and you lower the slide to 3, on a Chat that ends with the AI wraps up: the AI wraps up from the third message on, but the learner can still send two more. Raise the limit instead, and the code stops at the old limit before the AI wraps up.
The Storyline Web Object, Rise 360 and the HTML widget read the limit from SlideMind, so they follow it right away.
3. Edits to a published slide are live
Section titled “3. Edits to a published slide are live”A published slide has no separate draft. As the slide editor warns: “Saving updates the active configuration of this published slide.” The next learner who sends a message gets the new instructions, without republishing your course.
To change a slide that learners use:
- Test your unsaved changes in Try it first. It tests the unsaved draft and saves nothing.
- Save only when the result is right.
- If a change goes wrong, restore the earlier version from History. See Version history.
4. A Chat’s destination changes its replies right away
Section titled “4. A Chat’s destination changes its replies right away”SlideMind reads the slide’s destination at every Chat reply (or the workspace Default editor when the slide has none). A Chat whose destination is Rise 360 or the HTML widget answers in plain text, without its structured fields. If you switch a Chat installed with authoring tool code, or as a Storyline Web Object, to one of these destinations and save, the installed code stops receiving its fields at once, and its return variables stop updating. Switch back, or install the code of the new destination.