Flux 2 Flash Text To Image API Documentation

Flux 2 Flash Text To Image API Documentation

Playground

FLUX 2 flash is a fast, production-grade AI text-to-image generator from Black Forest Labs, optimized for realistic renders and crisp, typo-free text. It supports prompt-faithful styles and native editing—image-to-image, inpainting/outpainting, background replacement, and quick variants—for posters, logos, product shots, and social ads. Ready-to-use REST API with low latency, no cold starts, and affordable pricing.

Features

FLUX.2 Flash Text-to-Image is a fast, production-focused image generation model designed for high-volume, low-latency workflows. It turns a single text prompt into a ready-to-use image, making it a strong default for rapid iteration, batch pipelines, and “generate lots of options quickly” use cases.

This wrapper is text-to-image (text in, image out). If you need prompt-based edits to existing images (image-to-image, inpainting/outpainting, background replacement, etc.), use the related FLUX.2 Flash Edit model instead.

Key capabilities

• Fast, low-latency image generation for rapid iteration and large-scale creative production.
• Prompt-faithful composition and styles that hold up well across repeated runs and automation pipelines.
• Crisp on-image text tuned for cleaner typography (helpful for posters, product shots, and social creatives).
• Flexible output sizing via size or explicit width/height for banners, posters, square assets, and more.
• Reproducible variations with seed for controlled exploration and reruns.
• API-friendly output controls like synchronous waiting (enable_sync_mode) and base64 responses (enable_base64_output) for easier server-side integration.

Parameters and how to use

• prompt: (required) The text instruction that describes what you want to generate.
• size: A shorthand output size string (commonly formatted like WIDTH*HEIGHT).
• seed: Randomness control for reproducibility (-1 for a new random result each run).
• enable_sync_mode: If true, wait for generation/upload and return the result directly (API only).
• enable_base64_output: If true, return base64 output instead of a URL (API only).

Prompt

Write your prompt like you’re briefing a photographer or designer:

• Start with the subject + setting, then add style, camera/lighting, and details that matter (materials, mood, composition).

• For marketing/product visuals, include: background type, surface/reflection, lighting direction, and “clean” constraints (e.g., “no extra objects, no watermark”).

• For on-image text, keep it short and explicit:

• Put the exact text in quotes.

• Specify placement and typography cues (e.g., “centered headline, bold sans-serif, high contrast”).

• If you get typos, simplify the layout and reduce the amount of text.

Other parameters

• size Use when you want a quick preset-like size string. Across WaveSpeedAI FLUX endpoints, this is commonly written as WIDTH*HEIGHT (for example, 1024*1024). If you set width and height, keep them consistent with size to avoid ambiguity.

• width / height Use when you need exact dimensions (e.g., wide banners vs. tall posters).

• seed

• Use -1 for a fresh random result each run.

• Use a fixed integer (e.g., 12345) to reproduce a composition or generate controlled variations while you iterate on the prompt.

• enable_sync_mode (API only) Set to true when you want the call to wait and return the generated result in the same response (useful for simple backends and demos).

• enable_base64_output (API only) Set to true when you want the output encoded as base64 instead of a hosted URL (useful for storage, pipelines, or environments that can’t fetch URLs).

After you finish configuring the parameters, click Run, preview the result, and iterate if needed.

Pricing

$0.025 per run

Notes

• If you’re generating many variations, keep size (or width/height) and a fixed seed consistent while you tune the prompt—this makes changes easier to compare.

Related Models

• FLUX.2 Flash Edit – Prompt-based image editing (image-to-image) when you need to modify existing assets instead of generating from scratch.
• FLUX.2 Dev Text-to-Image – Leaner base model in the FLUX.2 family for fast iteration and training-friendly workflows.
• FLUX.2 Flex Text-to-Image – More style-forward generation when you want broader aesthetics while staying fast.
• FLUX.2 Pro Text-to-Image – Higher-capacity option for more demanding production visuals and higher fidelity outputs.

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",
  "size": "1024*1024"
}
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-2-flash/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