If you’re building a product that needs content at scale—short-form videos, cleaned-up images, and high-quality copy—you don’t want three separate services with three different auth schemes and job systems.

The SamsarOne Public API gives you a single base URL and a consistent workflow across:

• Video generation from text prompts or ordered image sequences
• Image operations like upscale/denoise and text removal for whitelabel assets
• Chat + embeddings for copy enhancement and semantic search over your own JSON data
• Assistant completions for OpenAI-compatible, session-aware responses

This post is a practical, developer-focused tour of the public endpoints, the common patterns (async jobs + status polling), and a few “full pipeline” examples you can adapt to your own app.

Quickstart: one request, one status poll​

Most video and image endpoints are asynchronous:

1. Submit a request and receive a request_id / session_id.
2. Poll GET /status?request_id=... until COMPLETED (or FAILED / CANCELLED).
3. Download the resulting asset from result_url (or result_urls). Completed video responses also include has_subtitles and result_language when session metadata is available.

Example: start a text-to-video job (subtitles enabled with a specific font):

curl -X POST https://api.samsar.one/v1/video/text_to_video \
  -H "Authorization: Bearer $SAMSAR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "input": {
      "prompt": "A timelapse of clouds over a futuristic city at sunrise",
      "duration": 30,
      "image_model": "GPTIMAGE2",
      "video_model": "RUNWAYML",
      "tone": "grounded",
      "aspect_ratio": "16:9",
      "language": "en",
      "enable_subtitles": true,
      "font_key": "Poppins"
    }
  }'

Then poll:

curl "https://api.samsar.one/v1/status?request_id=vid_a1b2c3d4" \
  -H "Authorization: Bearer $SAMSAR_API_KEY"

When complete, you’ll receive a result_url to the rendered video, plus video metadata such as has_subtitles and result_language.

The API at a glance​

• Base URL: https://api.samsar.one/v1
• Auth (recommended): Authorization: Bearer <API_KEY>
  • The API also accepts API_KEY: <API_KEY> in headers, and in some cases a bearer JWT user auth token.
• Content type: application/json
• Credits: most endpoints return usage headers when applicable:
  • x-credits-charged
  • x-credits-remaining

Core routes​

Video

• POST /video/text_to_video — generate a clip from a text prompt
• POST /video/image_list_to_video — generate a clip from an ordered image list (+ optional metadata)
• POST /video/upload_image_data — upload data URLs and get hosted image_urls back
• POST /video/add_outro_image — clone a session and append/replace an outro image layer
• POST /video/update_outro_image — update the outro image layer in a session clone (when an outro already exists)
• POST /video/translate_video — translate a completed session to a new language (creates a new session)
• POST /video/retranslate_video — alias for translating a completed session to a new language
• POST /video/join_videos — append multiple sessions into one (creates a new session)
• POST /video/clone and POST /v2/video/clone — deep-copy a completed session and render a new final video URL
• POST /video/add_subtitles — clone a session and add subtitle overlays before re-rendering
• POST /video/remove_subtitles — clone a session and remove text overlays before re-rendering
• POST /video/cancel_render — cancel an in-progress render for a session
• GET /video/fetch_latest_version — fetch the latest available render URL for a session_id

Image

• POST /image/assign_title — assign a short SEO-friendly image title from a file, data URL, or image URL (synchronous)
• POST /image/enhance — upscale/denoise
• POST /image/remove_branding — remove visible text from an image
• POST /image/add_image_set — save/extend an image set for downstream generation, including optional aspect_ratio
• POST /image/create_rollup_banner — generate a rollup banner from a tile list
• POST /image/enhance_and_generate_rollup_banner — (if needed) enhance low-res tiles, then generate a rollup banner
• GET /image/status — status for image jobs (accepts request_id or session_id)
• GET /image/list — list recent image sessions (history + current statuses)

Chat

• POST /chat/enhance — improve copy with optional metadata and language hints (synchronous)
• Embeddings:
  • POST /chat/create_embedding — create from JSON records or URL input
  • POST /chat/create_embedding_from_url — explicit URL-only create route
  • POST /chat/update_embedding
  • POST /chat/search_against_embedding(s)
  • POST /chat/similar_to_embedding(s)
  • POST /chat/delete_embedding / POST /chat/delete_embeddings
  • GET /chat/embedding_templates
  • GET /chat/embedding_status

Assistant

• POST /assistant/set_system_prompt — store or clear the account-level system prompt
• POST /assistant/completion — create a synchronous assistant response in modern OpenAI Responses format

Platform / billing / status

• GET /health — health check
• GET /supported_fonts — supported fonts by language code (no auth)
• GET /status — status for any video/image request id
• Credits + Stripe:
  • GET /credits
  • POST /credits/recharge
  • POST /enable_autorecharge
  • POST /auto_recharge/threshold
  • GET /payment_status (also supports POST)
• Client auth helpers:
  • POST /create_login_token (also supports GET)
  • GET /users/verify_token?loginToken=... (exchange login token for a long-lived client authToken)

Video: prompts, image sequences, and post-processing​

The Video API is built around a simple idea: enqueue a render job, then check status until a URL is ready.

Generate from text​

Use POST /video/text_to_video when you have a prompt and want the API to handle the rest. The request supports model selection, duration, aspect ratio, and subtitle settings:

{
  "input": {
    "prompt": "Required text, max 1000 chars",
    "image_model": "GPTIMAGE2 | NANOBANANAPRO | SEEDREAM",
    "video_model": "VEO3.1I2V | VEO3.1I2VFAST | COSMOS3SUPERI2V | SEEDANCEI2V | KLINGIMGTOVID3PRO | KLINGIMGTOVIDTURBO | HAPPYHORSEI2V | RUNWAYML",
    "duration": 30,
    "tone": "grounded | cinematic",
    "aspect_ratio": "16:9 | 9:16",
    "language": "auto | en | es | fr | de | pt | it | nl | sv | hi | ar | ru | ja | ko | th | zh",
    "enable_subtitles": true,
    "font_key": "Poppins"
  },
  "webhookUrl": "https://example.com/webhook"
}

Generate from an image list (story-first)​

If you already have images—product shots, frames, key visuals—POST /video/image_list_to_video creates a narrative and renders a clip from an ordered list:

• image_urls: list of URLs (strings, or objects containing image_url / enhanced_url)
• metadata: optional object that helps the narrative builder
• prompt: optional direction (“fast-paced teaser”, “grounded walkthrough”, etc.)
• video_model: optional; supports VEO3.1I2V, VEO3.1I2VFAST, COSMOS3SUPERI2V, SEEDANCEI2V, KLINGIMGTOVID3PRO, KLINGIMGTOVIDTURBO, HAPPYHORSEI2V, and RUNWAYML
• image_model: optional; supports GPTIMAGE2, NANOBANANAPRO, and SEEDREAM
• aspect_ratio: optional; supports 16:9 and 9:16
• backingtrack_model / music_provider: optional; choose LYRIA3 for Google Lyria 3 Pro native backing tracks or ELEVENLABS_MUSIC
• outro_image_url: optional final card or brand frame
• generate_outro_image: optional server-side CTA outro generation from the input images with either cta_url for a QR center image or outro_cta_image for a supplied center logo/CTA image
• add_footer_animation: optional per-scene footer QR section using footer_metadata
• limit_single_narrator: optional single narrator constraint for narration scenes
• add_narrator_avatar: optional influencer-style human narrator avatar overlay; automatically enables limit_single_narrator and adds 4 credits/sec

Use either a provided outro image or a generated CTA outro in a single request, not both. outro_cta_image uses { "top_text", "middle_image", "bottom_text" }.

Tip: if your images start life as base64/data URLs, first call POST /video/upload_image_data to get hosted URLs back.

Post-processing: outro, translate, join, remove subtitles​

Once you have a completed session, you can build “edit-like” workflows purely via API:

• Add or update an outro:
  • POST /video/add_outro_image appends (or replaces) an outro in a cloned session.
  • POST /video/update_outro_image updates the existing outro layer in a cloned session when an outro already exists.
• Translate a session: POST /video/translate_video (or POST /video/retranslate_video) creates a new session and runs the minimum required stages to output a translated video. The translate payload uses enable_subtitles (default false), translate_outro (default true), and translate_footer (default true); use the outro update routes when the image itself needs replacement.
• Join sessions: POST /video/join_videos appends multiple completed sessions into one new render (optionally blending scene boundaries).
• Deep clone a session: POST /video/clone or POST /v2/video/clone copies an owned completed session and queues only final video generation so the copy gets a new rendered video path and URL. This clone route is free.
• Add subtitles: POST /video/add_subtitles clones the session, adds text overlays, and rerenders frames/video.
• Remove subtitles: POST /video/remove_subtitles clones the session, strips text overlays, and rerenders frames/video.
• Cancel a render: POST /video/cancel_render cancels an in-progress render for a session id.
• Fetch the newest available render URL: GET /video/fetch_latest_version?session_id=... returns result_url, has_subtitles, and result_language when ready (or 202 while pending).

Image: enhance, whitelist, and build rollups​

The Image API is designed for production asset cleanup:

• Upscale/denoise (for low-res inputs that need to look crisp in motion)
• Remove visible text for marketplaces, partners, or repurposed creative
• Assign short SEO-friendly titles for image filenames or listing titles
• Manage image sets (save a batch once, then reuse downstream)
• Build rollup banners (tile-based promotional composites)

Generation and editing endpoints enqueue work; poll GET /image/status (or GET /status) using the returned request_id. POST /image/assign_title returns synchronously with { "content": "..." }.

A common pipeline: clean → enhance → rollup​

One pattern we see often:

1. Remove branding on a partner-supplied image with POST /image/remove_branding
2. Upscale with POST /image/enhance (pick resolution: 1k | 2k | 4k)
3. Generate a tile-based composite with POST /image/create_rollup_banner

If you want the system to automatically enhance low-res tiles before producing the composite, use:

• POST /image/enhance_and_generate_rollup_banner

You can also query history and current results with GET /image/list?limit=....

Chat: copy enhancement and embeddings for your data​

Chat endpoints are synchronous: you submit a request and get a response in the same call.

Enhance copy with metadata​

Use POST /chat/enhance to rewrite a message while keeping intent. You can provide metadata (audience, platform, tone) and a language hint:

{
  "message": "launching soon, stay tuned!",
  "metadata": { "channel": "instagram", "tone": "playful" },
  "language": "FR"
}

Embeddings: semantic search over your records or URLs​

Embeddings endpoints turn either a JSON array or a URL list into a searchable template:

• POST /chat/create_embedding — create a template from JSON records or urls
• POST /chat/create_embedding_from_url — explicit URL-only create route
• POST /chat/search_against_embedding — natural language query + optional structured filters
• POST /chat/similar_to_embedding — “show me more like this” from a query term or JSON payload
• POST /chat/update_embedding — upsert records later
• POST /chat/delete_embedding(s) — clean up by record id or wipe a template

This is a powerful building block for:

• product catalogs (“show me listings like this under $X”)
• internal docs (“find similar incident reports”)
• real estate, travel, marketplaces—any domain where you have semi-structured records plus free text

Credits and operational endpoints​

SamsarOne uses a credit-based billing model (1 USD = 100 credits). A few endpoints you’ll want in every integration:

• GET /credits — read your balance (and the latest top-up metadata)
• POST /credits/recharge — create a Stripe checkout link for a credit top-up
• POST /enable_autorecharge + POST /auto_recharge/threshold — keep production systems from stalling due to low balance
• GET /payment_status — poll Stripe checkout/setup completion
• GET /health — monitoring/uptime checks

You can also use POST /create_login_token to mint a short-lived login token that your client can exchange for a long-lived authToken via GET /users/verify_token.

Best practices (what we’ve learned from production integrations)​

• Keep API keys server-side. If you need client auth, mint a short-lived login token and exchange it for a client authToken.
• Treat request_id as your primary handle. Store it with your own job ids; poll GET /status until terminal states.
• Watch credit headers. Capture x-credits-charged / x-credits-remaining for usage tracking and alerting.
• Expect usage-based assistant pricing. Assistant requests scale with the amount of conversation context and answer generated, so larger multimodal requests will cost more than short text-only prompts.
• Handle 402 explicitly. Many endpoints return 402 when credits are insufficient—hook this into your top-up / auto-recharge flow.
• Use a webhook for video when you can. For long renders, webhooks reduce polling and make your system more event-driven.
• Send x-request-id for traceability. If you log request ids end-to-end, support and debugging get much easier.

Get started​

If you want to dive straight into the request/response schemas, start here:

• Generative API overview
• Video API
• Image API
• Chat API
• Assistant API
• Credits API
• Login Tokens

And if you prefer to try requests interactively first, the docs include a Playground for video, image, and chat workflows.

Docusaurus blogging features are powered by the blog plugin.

Here are a few tips you might find useful.

Blog posts support Docusaurus Markdown features, such as MDX.

This is the summary of a very long blog post,

Use a <!-- truncate --> comment to limit blog post size in the list view.

Lorem ipsum dolor sit amet...