Skip to content

ai infographic generator

An AI infographic generator you fix by drawing a box

Generate infographics, roadmaps, posters and cover art from a description, a template or a reference image — or from a research run, so the content holds up. To change something, box the area, say what it should become, and Divrite redraws only that part. Every version is kept.

  • Text or reference images
  • Edit several areas at once
  • Full version history
divrite · image · edit selected areas
1 2

annotations

  1. 1Make the centre label larger and bold
  2. 2Turn this segment into a bar chart

only the boxed areas change

versions

  1. v1
  2. v2
  3. v3

every version is kept · branch from any of them

Illustration

research first An image can be built on research too: Divrite researches the topic first and draws from the findings. How deep research works →

[01] real outputs

Visuals made with Divrite

The roadmap below came from a one-line request using a template. The pages next to it show how images sit inside larger pieces: charts for data, photos from the source pages, AI art for covers and concepts.

Every example here is a real Hot topics issue generated with Divrite.

See all Hot topics →

[02] how it works

Generate, then edit only what’s wrong

  1. step 01

    Describe

    Describe the image, start from a template, or add reference images to guide style and content. Pick portrait, landscape or square.

  2. step 02

    Generate

    Divrite works out what the image needs to say, plans the composition and where the text goes, then draws it.

  3. step 03

    Box what to change

    Drag boxes over the parts you want changed — as many as you need — and write what each should become. Everything outside the boxes stays the same.

  4. step 04

    Keep every version

    Each edit is a new version; earlier ones stay. Continue from any version, or describe a change to the whole image instead.

[03] vs. a chat assistant

Edits that don’t start from scratch

Content

Chat ·Drawn from the prompt alone.

Divrite ·Can be drawn from a research run, so the figures and labels come from real sources.

Edits

Chat ·Re-prompt and get a whole new image.

Divrite ·Box specific areas; only those are redrawn.

Several fixes

Chat ·One change per message.

Divrite ·Annotate several areas in one pass, each with its own instruction.

History

Chat ·Earlier results scroll away in the thread.

Divrite ·A version history you can branch from.

[$] pricing

From $9.9/mo

Every plan includes deep research and all output formats — presentations, reports, web pages and images. You pay with monthly credits, and each task shows its credit estimate before it runs. Final prices are shown at checkout.

View pricing

[04] faq

Frequently asked questions

What kinds of images work best?
Visuals that explain something: infographics, technology roadmaps, process diagrams, posters and cover images.
How does editing by annotation work?
Open the image, drag a box over each area you want changed and write what it should become. You can add an overall instruction and reference images too. Only the boxed areas are redrawn.
Can I use reference images?
Yes. Add reference images to guide style and content, both when generating and when editing. Each must be under 10 MB.
Are earlier versions kept?
Yes. Every edit creates a new version and the earlier ones stay, so you can compare them or continue from any of them.
Can Divrite research the topic before drawing?
Yes. Start with a research question and the image is generated once the research is done. For longer pieces, see deep research.
How are images chosen inside decks and reports?
By content: charts for data, photos from the source pages for real-world subjects, AI-generated art for covers and concepts, or your own images. No decorative filler. See AI presentations.

[05] related

More of what Divrite does

Make the visual, then fix only what’s wrong

Describe it, start from a template or start from research. Box what needs to change; keep everything else.