Abstract model routing architecture for the SJolt API platform

AI model API

SJolt: One API for the best generation models.

Explore models in the SJolt playground, compare output quality and cost, then move the same request flow into production.

PlaygroundTest models with API-ready inputs
CompareReview quality, cost, and status tracking
One APIUse one request flow across supported models
Usage billingTrack usage and cost from one balance

From model choice to production call

Move from model choice to production calls without switching tools.

With SJolt, compare generation and editing models in the catalog, test them with API-ready inputs, then copy the same request and result retrieval flow into your product.

Model output samples

Inspect outputs before choosing a model.

Explore models

Developer integration

Keep the same SJolt API request from playground to production.

TRY

The playground uses API request fields.

Prompt, ratio, duration, and source assets appear as request fields, so product experiments can move directly into engineering.

Open playground
COPY

Model pages include ready-to-use API examples.

Examples cover curl, JavaScript, Python, and task polling. Set credentials and model ID, then ship.

Read docs
ACCOUNT

One balance covers model usage.

Billing views show payment status, usage records, and the model-level costs needed for finance review.

View pricing

Operational notes

The questions teams ask before connecting generation models.

Model coverage, billing, async status tracking, and playground flow are visible before integration starts. Teams can judge fit before writing production code.

Which generation models does SJolt support?

SJolt supports generation workflows such as image generation, video generation, image-to-video, and image editing. The catalog organizes models by task, inputs, outputs, and status tracking.

Can I call different models through the same API?

Yes. SJolt uses one request pattern across supported generation models, with model-specific pricing and task parameters.

Do video generation jobs run asynchronously?

Yes. Long-running jobs run through async queues. Poll by task ID to retrieve status and results.

Can I test a model before integrating the API?

Yes. The playground uses the same fields as the API request, so prompts, ratios, durations, and model IDs can move directly into code.

Start with one model

Connect your first generation model with SJolt.

Pick a model, confirm output quality and cost in the playground, then reuse the API example in your product.

Start building