Flux Kontext Max Text To Image API Documentation

Flux Kontext Max Text To Image API Documentation

Playground

FLUX.1 Kontext [max] is a text-to-image model with max performance and greatly improved prompt adherence for accurate results. Ready-to-use REST inference API, best performance, no coldstarts, affordable pricing.

Features

FLUX Kontext Max is a high-quality text-to-image model built for prompt-faithful generation with strong composition control and cinematic detail. Write a clear scene description (subject, setting, action, camera, lighting, mood), and it produces polished, production-ready images—great for story frames, concept art, marketing visuals, and realistic or stylized illustration work.

Key capabilities

• High-fidelity text-to-image generation with strong prompt adherence
• Cinematic composition control (framing, lens feel, lighting, atmosphere)
• Handles complex scenes (multiple elements, detailed environments, coherent styling)
• Consistent output quality suitable for professional creative workflows

Pricing

$0.08 per image.

Total cost = num_images × $0.08 Example: num_images = 4 → $0.32

How to use

1. Write a prompt describing the scene and desired look.
2. Choose an aspect ratio for your target layout (e.g., 16:9 for cinematic, 1:1 for avatars).
3. Set a seed for repeatable results (optional), then generate.

Parameters

• prompt (required): The text description of what to generate
• seed: Fixed value for reproducibility; leave empty or random for variation
• aspect_ratio: Output aspect ratio (e.g., 16:9, 1:1, 9:16)
• enable_sync_mode: Wait for generation and return results directly (API only)

Prompting guide

A reliable prompt structure:

• Subject: who/what is in the image
• Action: what is happening
• Scene: where + time + atmosphere
• Camera: framing, lens vibe, movement (if relevant)
• Lighting: key light, contrast, color temperature
• Style: realistic, film still, illustration, etc.

Example pattern: A [shot type] of [subject] [action] in [scene]. [Lighting + mood]. [Camera/framing]. [Style cues].

Example prompts

• A cinematic close-up of a detective under flickering fluorescent lights, rain streaks on the window behind, shallow depth of field, tense atmosphere, film still look.
• Wide shot of a futuristic street market at dusk, neon reflections on wet pavement, dense crowd silhouettes, volumetric light beams, high detail, cinematic composition.
• Studio product photo of a minimalist smartwatch on a glossy black surface, softbox lighting, crisp reflections, premium advertising style, ultra clean background.

Best practices

• Keep the first sentence simple and concrete, then add constraints and style cues.
• Use camera and lighting language to steer composition more reliably than abstract adjectives.
• For consistent iteration, fix seed and adjust one variable at a time (prompt wording).

Authentication

For authentication details, please refer to the Authentication Guide.

API Endpoints

Submit Task & Query Result

set -euo pipefail

export WAVESPEED_API_KEY="your-api-key"

REQUEST_BODY=$(cat <<'JSON'
{
  "prompt": "A cinematic ocean wave at sunrise, highly detailed",
  "aspect_ratio": "1:1"
}
JSON
)

# 1. Submit the prediction once. Never retry this request automatically:
# a retry after an unclear failure can create a second, billed task.
CHECK='The task may already exist. Check History or the predictions list before submitting again.'
if ! SUBMIT_RESPONSE=$(curl --silent --show-error --fail-with-body \
  --connect-timeout 10 --max-time 60 \
  -X POST "https://api.wavespeed.ai/api/v3/wavespeed-ai/flux-kontext-max/text-to-image" \
  -H "Authorization: Bearer ${WAVESPEED_API_KEY}" \
  -H "Content-Type: application/json" \
  -d "${REQUEST_BODY}"); then
  printf 'Submit failed. %s\n' "${CHECK}" >&2
  exit 1
fi

TASK=$(printf '%s' "${SUBMIT_RESPONSE}" | jq 'if type == "object" and has("data") then .data else . end')
PREDICTION_ID=$(printf '%s' "${TASK}" | jq -r '.id // empty')
if [ -z "${PREDICTION_ID}" ]; then
  printf 'Submission response did not contain a prediction id. %s\n' "${CHECK}" >&2
  exit 1
fi
RESULT_URL="https://api.wavespeed.ai/api/v3/predictions/${PREDICTION_ID}/result"

# 2. Poll until the prediction reaches a final status.
BODY_FILE=$(mktemp)
trap 'rm -f "${BODY_FILE}"' EXIT
while true; do
  # Transient errors (429, 5xx) are retried, honoring Retry-After.
  HTTP_CODE=$(curl --silent --show-error --fail-with-body \
    --connect-timeout 10 --max-time 60 --retry 5 --retry-delay 2 \
    --output "${BODY_FILE}" --write-out '%{http_code}' \
    "${RESULT_URL}" \
    -H "Authorization: Bearer ${WAVESPEED_API_KEY}" || true)
  case "${HTTP_CODE}" in
    2??) ;;
    000|429|5??)  # dropped connection or temporary error: keep polling
      printf 'Task %s: HTTP %s, polling again in 10s\n' "${PREDICTION_ID}" "${HTTP_CODE}" >&2
      sleep 10; continue ;;
    *) printf 'Task %s: HTTP %s\n' "${PREDICTION_ID}" "${HTTP_CODE}" >&2
      cat "${BODY_FILE}" >&2; exit 1 ;;
  esac
  RESULT=$(jq 'if type == "object" and has("data") then .data else . end' "${BODY_FILE}")
  STATUS=$(printf '%s' "${RESULT}" | jq -r '.status // empty')

  case "${STATUS}" in
    completed) printf '%s\n' "${RESULT}" | jq '.outputs'; break ;;
    failed|cancelled|timeout|deleted)
      printf 'Task %s ended with status %s\n' "${PREDICTION_ID}" "${STATUS}" >&2
      printf '%s\n' "${RESULT}" | jq . >&2; exit 1 ;;
    *) sleep 2 ;;
  esac
done

Parameters

Task Submission Parameters

Request Parameters

Every model also accepts the global request options enable_sync_mode and enable_base64_output; see their pages for details.

Response Parameters

Result Request Parameters

Result Response Parameters