Skip to content

api-reference.md

Latest commit

 

History

History
375 lines (290 loc) · 9.28 KB

api-reference.md

File metadata and controls

375 lines (290 loc) · 9.28 KB

deAPI API Reference

Base URL

https://api.deapi.ai/api/v1/client/

Authentication

All endpoints require Bearer token authentication:

Authorization: Bearer $DEAPI_API_KEY

Endpoints

Text-to-Image (txt2img)

Generate images from text prompts.

Request:

POST /txt2img
Content-Type: application/json

{
  "prompt": "a sunset over mountains, photorealistic",
  "model": "Flux1schnell",
  "width": 1024,
  "height": 1024,
  "steps": 4,
  "guidance": 7.5,
  "seed": 123456,
  "negative_prompt": "blurry, low quality"
}

Parameters:

Parameter Type Required Default Description
prompt string Yes - Image description
model string Yes - Flux_2_Klein_4B_BF16, Flux1schnell, or ZImageTurbo_INT8
width int Yes 1024 Image width (256-2048, depends on model)
height int Yes 1024 Image height (256-2048, depends on model)
steps int Yes 4 Inference steps (model-dependent)
guidance float No 7.5 Prompt adherence (not supported by Klein)
seed int Yes - Random seed (0-999999)
negative_prompt string No - What to avoid (not supported by Klein)

Models:

Model Slug Steps Max size Guidance Negative prompt
FLUX.2 Klein 4B Flux_2_Klein_4B_BF16 4 (fixed) 1536 (step: 16) No No
Flux.1 Schnell Flux1schnell 1-10 (default: 4) 2048 (step: 128) No Yes
Z-Image-Turbo ZImageTurbo_INT8 1-50 (default: 8) 2048 (step: 16) No Yes

Image-to-Image (img2img)

Transform existing images with style prompts.

Request:

POST /img2img
Content-Type: multipart/form-data

-F "image=@photo.jpg"
-F "prompt=convert to watercolor painting"
-F "model=Flux_2_Klein_4B_BF16"
-F "steps=4"
-F "seed=123456"

Parameters:

Parameter Type Required Default Description
image file Yes - Source image file(s) (max 10MB). Klein supports up to 3 images
prompt string Yes - Transformation description
model string Yes - Flux_2_Klein_4B_BF16 or QwenImageEdit_Plus_NF4
guidance float No 7.5 Prompt adherence (Klein: not supported, Qwen: 1-20)
steps int Yes - Inference steps (Klein: 4 fixed, Qwen: 1-50 default 40)
seed int Yes - Random seed (0-999999)
width int No - Output width (Klein only, 256-1536, step: 16)
height int No - Output height (Klein only, 256-1536, step: 16)

Models:

Model Slug Steps Max images Guidance Custom output size
FLUX.2 Klein 4B Flux_2_Klein_4B_BF16 4 (fixed) 3 No Yes (256-1536)
Qwen Image Edit Plus QwenImageEdit_Plus_NF4 1-50 (default: 40) 1 Yes No

Text-to-Audio (txt2audio)

Convert text to speech.

Request:

POST /txt2audio
Content-Type: application/json

{
  "text": "Hello, welcome to our service.",
  "voice": "af_bella",
  "model": "Kokoro",
  "lang": "en-us",
  "speed": 1.0,
  "format": "mp3",
  "sample_rate": 24000
}

Parameters:

Parameter Type Required Default Description
text string Yes - Text to convert
voice string Yes af_bella Voice ID (54+ available)
model string Yes - Kokoro
lang string Yes - Language code: en-us, en-gb, ja, zh, etc.
speed float Yes 1.0 Speech speed (0.5-2.0)
format string Yes mp3 Output: mp3, wav, flac, ogg
sample_rate int Yes 24000 Sample rate: 22050, 24000, 44100, 48000

Video Transcription (vid2txt)

Transcribe video content from URL.

Request:

POST /vid2txt
Content-Type: application/json

{
  "video_url": "https://youtube.com/watch?v=xxx",
  "include_ts": true,
  "model": "WhisperLargeV3"
}

Parameters:

Parameter Type Required Default Description
video_url string Yes - Video URL (YouTube, Twitch, Kick, X)
include_ts bool Yes false Include timestamps
model string No WhisperLargeV3 Whisper model

Audio Transcription (aud2txt)

Transcribe audio from URL (X Spaces, etc.).

Request:

POST /aud2txt
Content-Type: application/json

{
  "audio_url": "https://example.com/audio.mp3",
  "include_ts": true,
  "model": "WhisperLargeV3"
}

Parameters:

Parameter Type Required Default Description
audio_url string Yes - Audio URL
include_ts bool Yes false Include timestamps
model string No WhisperLargeV3 Whisper model

OCR (img2txt)

Extract text from images.

Request:

POST /img2txt
Content-Type: multipart/form-data

-F "image=@document.png"
-F "model=Nanonets_Ocr_S_F16"

Parameters:

Parameter Type Required Default Description
image file Yes - Image file (max 10MB)
model string Yes - Nanonets_Ocr_S_F16

Background Removal (img-rmbg)

Remove image backgrounds.

Request:

POST /img-rmbg
Content-Type: multipart/form-data

-F "image=@photo.jpg"
-F "model=Ben2"

Parameters:

Parameter Type Required Default Description
image file Yes - Image file (max 10MB)
model string Yes - Ben2

Output: PNG with transparent background

Image Upscaling (img-upscale)

Increase image resolution.

Request:

POST /img-upscale
Content-Type: multipart/form-data

-F "image=@photo.jpg"
-F "model=RealESRGAN_x4"

Parameters:

Parameter Type Required Default Description
image file Yes - Image file
model string Yes - RealESRGAN_x2 (2x) or RealESRGAN_x4 (4x)

Text-to-Video (txt2video)

Generate videos from text.

Request:

POST /txt2video
Content-Type: multipart/form-data

-F "prompt=a cat walking through a garden"
-F "model=Ltxv_13B_0_9_8_Distilled_FP8"
-F "width=768"
-F "height=512"
-F "guidance=7.5"
-F "steps=30"
-F "frames=120"
-F "seed=123456"

Parameters:

Parameter Type Required Default Description
prompt string Yes - Video description
model string Yes - Ltxv_13B_0_9_8_Distilled_FP8
width int Yes 768 Video width
height int Yes 512 Video height
guidance float Yes 7.5 Prompt adherence
steps int Yes 30 Inference steps
frames int Yes 120 Number of frames
seed int Yes - Random seed
fps int No 24 Frames per second

Image-to-Video (img2video)

Animate static images.

Request:

POST /img2video
Content-Type: multipart/form-data

-F "first_frame_image=@photo.jpg"
-F "prompt=gentle movement, cinematic"
-F "model=..."
-F "width=768"
-F "height=512"
-F "guidance=7.5"
-F "steps=30"
-F "frames=120"
-F "seed=123456"

Parameters:

Parameter Type Required Default Description
first_frame_image file Yes - Source image file
prompt string Yes - Motion description
model string Yes - Model name
width int Yes - Video width
height int Yes - Video height
guidance float Yes 7.5 Prompt adherence
steps int Yes 30 Inference steps
frames int Yes 120 Number of frames
seed int Yes - Random seed

Text Embeddings (txt2embedding)

Generate vector embeddings.

Request:

POST /txt2embedding
Content-Type: application/json

{
  "input": "Your text to embed",
  "model": "Bge_M3_FP16"
}

Parameters:

Parameter Type Required Default Description
input string/array Yes - Text to embed (single or array)
model string Yes - Bge_M3_FP16

Request Status (request-status)

Check job status.

Request:

GET /request-status/{request_id}

Response:

{
  "request_id": "abc123",
  "status": "done",
  "result_url": "https://...",
  "created_at": "2024-01-15T10:30:00Z",
  "completed_at": "2024-01-15T10:30:45Z"
}

Status values:

  • processing - Job in progress
  • done - Job completed successfully
  • failed - Job failed (check error field)

Result Delivery Methods

Method Best for Setup
Polling CLI, scripts Default, no config needed
Webhooks Backend, serverless Add webhook_url to requests
WebSockets Real-time UI Pusher connection

See webhooks docs and websockets docs.