Skip to content

Multi-page generation guide

Create multiple comic pages in one bounded API request.

Multi-page mode uses pagination.totalPages and pagination.panelsPerPage together so one create request describes a page grid instead of a single page with fixPanelNum.

Pagination defines requested structure, not guaranteed pacing, continuity, lettering quality, or publication readiness. Review each returned page and keep story planning outside the generation request.

Inspect the pagination contract

Account, API key, usage, and billing open on LlamaGen.AI.

Direct answer

The decision this page is for.

Last verified:

To create a multi-page comic, send pagination.totalPages and pagination.panelsPerPage together in the initial POST. Do not send fixPanelNum in the same request. The page grid bounds the job, but the returned pages still require pacing, continuity, lettering, and publication review.

Verifiable request

Create two pages with four panels each

cURL
curl -X POST \
  https://api.llamagen.ai/v1/comics/generations \
  -H "Authorization: Bearer $LLAMAGEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "prompt": "An original harbor courier mystery told across two pages. End page one on a discovery and resolve the delivery on page two.",
  "pagination": {
    "totalPages": 2,
    "panelsPerPage": 4
  },
  "language": "English"
}'

Structured response

State your application can route

response.jsonILLUSTRATIVE
{
  "id": "gen_abc123",
  "status": "LOADING"
}

Pagination walkthrough

Review the whole page batch, not only the first image.

The working table connects a bounded page grid, consecutive story beats, recurring characters, and an editor-led acceptance pass.

An independent comics editor reviews four connected storm-rescue page proofs beside a blank multi-page thumbnail plan
Editorial illustration generated with AI; it is not a creator photograph or customer work and not a generated API result.Art direction · American Sunday newspaper production documentary

Page-batch workflow

Plan the story beats before choosing the page grid.

A multi-page request is most useful when the application knows what each page must accomplish and can stop the batch for review before release.

  1. 01Writer or product

    Scope the batch

    Define the story beat, ending condition, required characters, and review owner for this page group.

  2. 02Application

    Choose the grid

    Set totalPages and panelsPerPage together; do not send fixPanelNum in the same request.

  3. 03Server

    Create once

    Submit the bounded request from a trusted runtime and persist the generation ID and requested pagination.

  4. 04Worker

    Track the job

    Show the documented asynchronous state instead of keeping one browser request open.

  5. 05Editor

    Review by page

    Check story order, continuity, captions, panel count, and transitions across every returned page.

  6. 06Publishing system

    Archive the accepted batch

    Store accepted assets and inputs before continuation, selective regeneration, lettering, or prepress.

Capability and boundary facts

Facts an integration can verify.

Last verified:

Operation
POST /v1/comics/generations
Multi-page fields
pagination.totalPages and pagination.panelsPerPage
Mutual exclusion
Do not combine pagination with fixPanelNum
Illustrative response
Generation ID and lifecycle status only; not live account output
Verification
LlamaGen human documentation and OpenAPI checked on 2026-10-29

Good fit

Use it when

  • Bounded issue or episode batches
  • Comic-book page prototypes
  • Editorial workflows that review pages separately

Wrong fit

Do not use it for

  • Combining pagination with fixPanelNum
  • Generating an uncontrolled full series
  • Treating returned pages as final lettering or print files

Production limits

Plan for the boundaries.

  • 01The current human documentation lists 1–20 for each pagination field, while the current OpenAPI contract declares the minimum but not those maximums. Re-check both sources before hard-coding a maximum.
  • 02The same sources currently disagree on the maximum for fixPanelNum: the human guide says 20 and OpenAPI says 40. This page uses pagination and does not resolve that separate contract discrepancy.
  • 03A requested page grid does not guarantee scene pacing, character continuity, panel composition, or a fixed completion time.
  • 04The response is contract-shaped illustrative data; no paid generation was executed for this page.
  • 05Generation is asynchronous; production interfaces must handle processing and failure states.
  • 06Outputs should be reviewed against product, brand, safety, and print requirements before delivery.
  • 07You must have the rights to prompts, photos, scripts, and reference assets submitted to the API.

Questions

Direct answers before you integrate.

01Which fields create a multi-page comic?
Use the nested pagination object with both totalPages and panelsPerPage. The current contract rejects combining pagination with fixPanelNum, and top-level totalPages or panelsPerPage are not the documented request shape.
02What is the maximum page or panel value?
The LlamaGen human documentation reviewed on 2026-10-29 lists 1–20 for each pagination field, while OpenAPI declares a minimum of 1 without those maximums. Validate conservatively and re-check both sources before hard-coding a maximum.
03Why does this page not quote one fixPanelNum maximum?
The current human guide says 20 and OpenAPI says 40. That source discrepancy does not affect the conservative pagination example here, but it should remain visible until LlamaGen aligns the contracts.
04Does pagination guarantee a finished issue?
No. It requests a page and panel structure; it does not guarantee narrative pacing, character continuity, readable lettering, print readiness, or deterministic completion. The example response is contract-shaped, not a live paid result.

Next step

Move from evidence to a real API request.

LlamaGen provides the playground, API key, usage, billing, and authoritative service documentation.

Inspect the pagination contract