> ## Documentation Index
> Fetch the complete documentation index at: https://apidocs.overtenai.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Slides guide

> Generate .pptx decks — pitch decks, quarterly reviews, technical talks.

The Slides endpoint produces professional `.pptx` output with a cohesive
visual system: one palette across the deck, consistent type scale, visual
pacing between hero / content / section-break slides. The agent thinks
like a senior UI/UX designer.

## What the Slides agent can do

Ordered from "used in almost every deck" to specialized:

* **Multi-slide decks** with a cohesive theme — the agent picks a
  color palette, accent color, type scale, and spacing system once and
  applies it consistently across every slide.
* **Rich layouts** — hero / title slides, agenda slides, two-column
  and three-column content, KPI / big-number cards, comparison
  matrices, quote breaks, thank-you closers. The agent picks the right
  layout per slide based on what the content is doing.
* **Data visualizations** — column / bar / line / pie / combo /
  donut / scatter / timeline / roadmap charts with the deck's accent
  color; KPI cards with prominent numbers; process flowcharts and
  pyramid / funnel diagrams.
* **Tables on slides** — headed tables, shaded rows, per-cell borders
  and alignment; great for feature comparisons and metrics breakdowns.
* **Text styling per shape** — font family, size, weight, color,
  alignment, line height; bold and italic runs inline; bulleted and
  numbered lists.
* **Backgrounds** — solid color, gradient (linear or radial), or
  uploaded image; per-slide override so hero and section-break slides
  can differ from content slides.
* **Speaker notes** — the agent writes presenter notes per slide when
  the prompt asks for them; notes travel with the `.pptx` and appear
  in PowerPoint / Keynote / Google Slides presenter view.
* **Images and icons** — embed uploaded assets sized and positioned to
  the layout; re-use the same image across multiple slides.
* **Shapes and callouts** — rectangles, circles, arrows, connectors,
  process-flow shapes — positioned precisely with x/y/width/height.
* **Slide-level actions** — insert, replace, duplicate, delete, or
  reorder specific slides in an existing deck (via follow-up calls
  with the same `run_id`).
* **Outline-first drafting** — produce just the slide titles and
  bullets via `POST /slides/outline` for preview / approval before
  committing to a full deck build.

Describe the deck's purpose and audience; the agent picks the layouts,
palette, and pacing. You can steer with explicit hints ("dark-navy
theme", "Series B pitch deck", "include a timeline on slide 7") and
the agent respects them.

## Quickstart

```python theme={null}
resp = requests.post(
    f"{API}/slides/generate",
    headers={"X-API-Key": KEY},
    json={
        "prompt": (
            "10-slide Q3 investor update for our SaaS company. "
            "Focus: MRR growth, churn reduction, Q4 roadmap. "
            "Visual style: clean, minimal, dark navy accents."
        ),
    },
)
```

Output: a 10-slide `.pptx` with hero opener, content slides, section
breaks, and a closing CTA.

## Presentation modes

The agent picks the right mode from the prompt, but you can nudge it:

| Mode                       | Signal words                                           | Typical output                                         |
| -------------------------- | ------------------------------------------------------ | ------------------------------------------------------ |
| **Visual-first / minimal** | "pitch deck", "narrative", "storytelling", "clean"     | Big hero images, few bullets, emphasis on hierarchy    |
| **Balanced**               | "investor update", "business review"                   | Mix of visual and data; moderate text density          |
| **Information-dense**      | "technical deep-dive", "detailed analysis", "academic" | Tighter layouts, more tables/charts, appendix-friendly |

<Tip>
  For pitch decks, keynotes, and executive summaries, the agent defaults
  to **visual-first**. For internal reporting and technical content, it
  leans **information-dense**.
</Tip>

## Working with images

Image generation is NOT part of the API. Upload your own and reference
them:

```python theme={null}
hero = requests.post(
    f"{API}/images", headers=headers,
    files={"file": open("hero.png", "rb")},
    data={"caption": "Q3 hero", "tags": '["hero"]'},
).json()

product = requests.post(
    f"{API}/images", headers=headers,
    files={"file": open("dashboard.png", "rb")},
    data={"tags": '["product", "screenshot"]'},
).json()

resp = requests.post(f"{API}/slides/generate", headers=headers, json={
    "prompt": "10-slide pitch deck with product screenshots",
    "image_assets": [
        {"image_id": hero["image_id"], "tags": ["hero"]},
        {"image_id": product["image_id"], "tags": ["product"]},
    ],
})
```

If your prompt references imagery but no `image_assets` are supplied,
[preflight](/concepts/preflight) rejects the request upfront with
`400 preflight_failed: missing=["images"]`.

## Brand customization

Pass a `brand` object to enforce your visual identity:

```python theme={null}
resp = requests.post(f"{API}/slides/generate", headers=headers, json={
    "prompt": "10-slide pitch deck on Series B round",
    "brand": {
        "primary_color": "#2E75B6",
        "font": "Inter",
        "logo_url": "https://cdn.acme.com/logo.svg",
    },
    "image_assets": [...],
})
```

The agent threads the color through headers, accents, dividers, and the
closing CTA. Font applies to all slide text.

## Editing existing decks

```python theme={null}
upload = requests.post(
    f"{API}/files", headers=headers,
    files={"file": open("deck_draft.pptx", "rb")},
    data={"purpose": "init"},
).json()

resp = requests.post(f"{API}/slides/generate", headers=headers, json={
    "prompt": "Replace slide 3 with a KPI summary showing three metric cards",
    "init_file": upload["file_id"],
})
```

## Outline-only generation

If you want the agent to draft just the slide structure (titles + bullets)
before committing to a full deck:

```python theme={null}
resp = requests.post(f"{API}/slides/outline", headers=headers, json={
    "prompt": "10-slide Series B pitch deck",
    "slide_count_hint": 10,
})
# → {"outline": [{"title": "...", "bullets": [...]}, ...]}
```

Use this for preview / approval workflows. Once the outline's good, pass
the same `run_id` to `/slides/generate` with a follow-up prompt like "now
build the full deck from this outline".

## Slide types

The agent builds three canonical types:

* **Hero** — dark / accent background, bold headline. Opening and closing
  slides default to this.
* **Content** — white/light background, 3–5 bullets max, consistent type.
* **Section break** — gradient tint, big section title, transitions
  between major chapters.

Visual pacing: no more than 2 consecutive slides with the same layout family.

## Things the agent handles well

* **Cohesive palette** — one industry preset (finance, tech, healthcare,
  etc.) per deck. Inherited by every slide.
* **Chart-led proof slides** — market size, traction, comparison matrices.
* **Split-panel layouts** — image left, text right (or vice versa).
* **Sticky type scale** — headline, body, caption — applied uniformly.
* **Contrast rules** — text on dark backgrounds is always light, and
  vice versa. Never invisible.

## Things to be aware of

* **Image search is not available in API mode.** Upload what you want to
  use. The agent won't search the web for images.
* **Max slide count: 30 per call.** Larger decks should be built in
  chunks via `run_id` resume — each follow-up call costs only a
  fraction of the first because the agent remembers the deck's style.
* **Speaker notes** are supported — the agent writes them into the
  notes pane of each slide when your prompt asks for them (e.g. "add
  speaker notes under each slide for the presenter"). They travel with
  the `.pptx` and show up in PowerPoint / Keynote / Google Slides
  presenter view.
* **Pattern selection is agent-driven.** Describe the layout you want
  in your prompt ("use a three-column comparison on slide 4", "make
  slide 7 a big-number KPI card"); the agent matches to the closest
  supported pattern.