REST API

API documentation

Video generation, speech synthesis, image generation and voice cloning. One API key, one base URL: https://seslab.com.tr/api/v1

Get a key Open in the panel

Getting started (3 steps)

Get your API key

Open an account and create a new key from the “My keys” page in the panel. Your key starts with sk_ (e.g. sk_a1b2c3…) and is shown only once — copy it and store it safely.

Send your first request

Add the Authorization: Bearer sk_… header to every request and send the data as JSON. The easiest first request: GET /me — it returns your account and balance.

Collect the result

Images and video can take a while: you first get an id, then ask for the status with that id. Credits are charged on success and refunded automatically on failure.

Authentication

Add the Authorization: Bearer sk_… header to every request. You create the key in the panel; an invalid or missing key returns 401, and too many failed attempts return 429. There is no per-minute request limit on a valid key.

curl https://seslab.com.tr/api/v1/me \
  -H "Authorization: Bearer sk_YOUR_API_KEY"

Desktop app token

Rule: sk_ is used only for the exchange; it is never stored permanently inside the desktop app. The desktop app uses the sk_ API key only on the first exchange request. Store the dsk_ token from the POST /api/desktop/auth/token response securely in the macOS Keychain or Windows Credential Manager. Only a hash of that token is kept in the SesLab database.

curl -X POST https://seslab.com.tr/api/desktop/auth/token \
  -H "Authorization: Bearer sk_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"device_label":"Ufuk Mac M4","app_version":"1.0.0"}'

# Response: { "desktop_token": "dsk_...", "expires_in": 2592000 }

Endpoints

POST

/videoStart a video generation (async, 202 + id)by tier

GET

/video?id=IDQuery a video’s status / resultfree

POST

/ttsTurn text into speech (sync if ready in ≤45 s, otherwise 202 + id)1.0 cr/char · v3: 1.5 cr/char

GET

/tts?id=IDQuery a TTS job’s status / audiofree

POST

/api/image-generateStart an image generation (async, 202 + id)by model

GET

/api/image-generate/status?id=IDQuery a long-running image jobfree

GET

/voicesAvailable voice cataloguefree

POST

/voice-cloneCreate a voice clone (multipart/form-data)free

GET

/voice-cloneList your voice clones and their voice_idsfree

DELETE

/voice-clone?voice_id=IDDelete a voice clonefree

GET

/meAccount details and balancefree

GET

/creditsCredit balancefree

GET

/statusSystem status (no auth required)free

POST

/desktop/auth/tokenExchange an sk_ key for a short-lived dsk_ token for the desktop appfree

Video generation

POST /video

Returns 202 (async). Poll the result with GET /video?id=. Credits are refunded automatically on failed jobs.

Field	Type	State	Description
prompt	string (10–1000)	required	Describe what you want to see in the video. English prompts usually give better results.
model	string	opt. · def. standard-video	Video package: standard-video (Fast) or hd-video (Premium). Leave it empty and standard-video is used. Old package names have been removed; if you send one you get a clear error telling you what to use instead.
resolution	string	opt. · def. 720p	Resolution: 720p (default) or 1080p. 1080p costs more credits within the same package.
aspect_ratio	string	opt. · def. 16:9	Only 16:9 landscape video is supported.
duration	number	opt. · def. 8	The duration is fixed: every video package produces 8 s, and the value you send is ignored.

Quality tiers and credits

Tier	model =	Credits	Notes
Fast default	`standard-video`	19,500	16:9 · 720p · 8 s
Fast (1080p)	`standard-video`	29,250	16:9 · 1080p · 8 s · send resolution=1080p
Premium	`hd-video`	25,500	16:9 · 720p · 8 s
Premium (1080p)	`hd-video`	38,250	16:9 · 1080p · 8 s · send resolution=1080p

Note: the video endpoint currently supports 16:9 landscape only. If aspect_ratio is left empty, 16:9 is used.

Speech synthesis (TTS)

POST /tts

If it is ready within 45 seconds you get a synchronous audio_url. Long texts return 202 + id; poll with GET /tts?id=. For ElevenLabs send one of the voice-natural packages, or voice-economy as the budget option.

Field	Type	State	Description
text	string (1–31000)	required	The text to turn into speech.
voice_id	string	required	Which voice reads it. Fetch the voice list with GET /voices and put the id from there here.
voice_package	string	opt. · def. voice-natural	ElevenLabs: voice-natural, voice-natural-fast, voice-natural-live or voice-natural-latest. Budget option: voice-economy.
language	string	optional	Language of the text, e.g. tr (Turkish). Only valid for voice-natural and voice-fast.
voice_settings	object	optional	Fine tuning (voice-natural only): stability 0–100, similarity_boost 0–100, speed 0.7–1.2, style 0–100, use_speaker_boost: true/false.

Image generation

POST /api/image-generate

A SesLab-compatible endpoint. Only the verified active image model is open right now. Returns 202 + id; poll the result with the status endpoint. Credits are refunded automatically on failure.

Field	Type	State	Description
prompt	string (1–20000)	required	What should be in the image? Describe it freely (Turkish works too).
model	string	required	Model ID: only nano-banana-2-lite is active right now. image-standard is accepted as a legacy alias.
aspect_ratio	string	optional	Aspect-ratio choice for the active model.
resolution	string	disabled	The active model runs natively at 1K; extra resolution choices are disabled.
mode	string	disabled	Image services that are not active are disabled.
quality	string	disabled	Image services that are not active are disabled.
image_input	string[]	optional	Reference images: base64 data URI or http(s) URL. Alias: reference_images. Up to 4 on the active model.
negative_prompt	string	optional	Things to avoid.

Image models

nano-banana-2-liteNano Banana 2 Lite2,000 crverified active model · 1K · 4 references

default

GET /api/image-generate/status?id=ID

Returns the status, image_url, image_urls and error fields for long-running image jobs.

Voice cloning

POST /voice-clone — multipart/form-data

Free. Returns a voice_id — use it in TTS calls.

Field	Type	State	Description
file	file	required	The voice sample to clone — MP3/M4A/WAV/OGG/FLAC, max 10 MB (10 seconds to 5 minutes of clean audio).
voice_name	string (≤100)	required	The name you give the clone (visible to you).
language_tag	string	opt. · def. Turkish	Language of the voice, e.g. Turkish.
gender_tag	string	opt. · def. female	Gender of the voice: female or male.

GET /voice-clone

Lists every voice clone on your account, newest first, with their voice_id fields. Takes no parameters.

{
  "success": true,
  "voices": [
    {
      "id": "b7c1e0a2-4f10-4a91-9c2e-1d5b7a0e33af",
      "voice_id": "2630054",
      "cloned_voice_id": "2630054",
      "name": "My voice",
      "created_at": "2026-08-12T09:14:22.000Z"
    }
  ]
}

Speaking with a clone: send the voice_id (or id) you get here as voice_id in your POST /tts call — no package/model field is needed, and it is ignored for clone voices. The text limit is 5,000 characters.

DELETE /voice-clone?voice_id=ID

Deletes the clone. The voice_id query parameter is required; if the clone is not found you get a 404.

Code samples

# Basic auth check
curl https://seslab.com.tr/api/v1/me \
  -H "Authorization: Bearer sk_YOUR_API_KEY"

# Start a video — packages: standard-video (Fast) / hd-video (Premium), 16:9, 720p|1080p, 8 s
curl -X POST https://seslab.com.tr/api/v1/video \
  -H "Authorization: Bearer sk_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "Cinematic product showcase, smooth camera move",
    "aspect_ratio": "16:9",
    "model": "standard-video",
    "resolution": "720p"
  }'

# Query the video status (free, poll repeatedly)
curl "https://seslab.com.tr/api/v1/video?id=VIDEO_ID" \
  -H "Authorization: Bearer sk_YOUR_API_KEY"

# Turn text into speech
curl -X POST https://seslab.com.tr/api/v1/tts \
  -H "Authorization: Bearer sk_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "text": "Hello, this voice was generated with Seslab.",
    "voice_id": "VOICE_ID",
    "voice_package": "voice-natural",
    "language": "en"
  }'

# Generate an image — returns 202 + id, poll the result from the status endpoint
curl -X POST https://seslab.com.tr/api/image-generate \
  -H "Authorization: Bearer sk_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "Premium product photo with blue lighting on a dark background",
    "aspect_ratio": "1:1",
    "model": "nano-banana-2-lite"
  }'

Error codes

401UnauthorizedThe Authorization header is missing or the key is wrong.

402Insufficient CreditsNot enough credits. Top up from the panel.

422Validation ErrorMalformed request body or unsupported value.

429Too Many Attempts10+ failed authentication attempts in a minute.

404Not FoundThe requested generation (id) was not found.

503Service Not ReadyThe generation service is not active right now.

500Server ErrorUnexpected error — try again; credits are refunded on a failed job.

Ready to start?

Open an account, create your API key and send your first request within minutes.

Start free Pricing