Text Generator

Official API verified 2026-07-31

Generate text with supported Gemini models through synchronous, asynchronous, or Server-Sent Events streaming workflows.

Endpoints

Operation	Endpoint
Create sync or async task	POST /sse/text_generator/
Stream a task	POST /sse/text_generator/stream/
List tasks	GET /api/text_generator/
Get task	GET /api/text_generator/{id}/

Required input

• model_name
• messages, using user or model roles

Optional fields include system_instruction, generation config, tools, tool configuration, webhook_url, and sync.

Choose a mode

Mode	Behavior	Best use
Sync	Set sync: true; request waits for the completed response	Short deterministic jobs
Async	Default; returns pending	Longer or multimodal jobs, polling, and webhooks
Streaming	Returns Gemini-native chunks over SSE	Interactive text output

Async example

curl --request POST \
  --url https://api.creatify.ai/sse/text_generator/ \
  --header 'Content-Type: application/json' \
  --header "X-API-ID: $CREATIFY_API_ID" \
  --header "X-API-KEY: $CREATIFY_API_KEY" \
  --data '{
    "model_name": "gemini-2.5-flash",
    "messages": [
      {"role": "user", "content": "Write three source-backed hooks for a local SEO and GEO video."}
    ],
    "system_instruction": "Return concise marketing hooks without unsupported claims."
  }'

Poll GET /api/text_generator/{id}/ until status is done or use webhook_url. Read response_text, response_function_calls, failed_reason, and credits_used from the result.

Streaming boundary

The streaming endpoint returns text/event-stream. Its final SSE chunk includes a Creatify object containing the generation ID and credits used.

Safety and cost

• Keep API credentials server-side.
• Treat HTTP 402 as insufficient credits and HTTP 429 as a rate-limit response.
• Use async mode for image or video inputs and other potentially slow requests.
• Record the returned credits_used rather than estimating text-generation cost.

Source: official Text Generator overview.