1 / 3
【Overview】 SJolt is a single API and playground that aggregates image, video, audio, and text generation models. It solves the problem of juggling multiple providers by offering one request pattern across supported models, so teams can compare output and cost in a playground before moving the same request flow into production. SJolt is aimed at product teams, developers, and creators who need to evaluate and integrate generative models without rebuilding their pipeline for each one. The playground uses the same fields as the API request, so experiments translate directly into code. Long-running jobs are handled through async queues, and users can poll by task ID for status and results. Webhook callbacks notify a top-level URL when a task reaches a terminal state, and usage-based billing tracks model usage and cost against a single account balance. SJolt also includes Seedance 2.5 text-to-video, which creates videos up to 30 seconds combining up to 50 multimodal references with frame-level control. A 480p-draft preview can be upscaled to 1080p for final output. The SJolt MCP connects Claude or ChatGPT to the remote MCP server for model access. 【Key features】 • Unified generation API: One request pattern across supported image, video, audio, and text generation models with model-specific parameters. • Playground testing: Test models with the same fields used in the API request so experiments move directly into code. • Async task handling: Runs long-running generation jobs through async queues and lets you poll by task ID for status and results. • Webhook callbacks: Sends a POST to a top-level webhook URL when a task reaches a terminal state. • Usage-based billing: Tracks model usage and cost against a single account balance. • SJolt MCP: Connects Claude or ChatGPT to the sjolt remote MCP server. 【Use cases】 1. text-to-video generation 2. image-to-video generation 3. reference-to-video generation 【Pros and cons】 Pros: - Developers benefit from a consistent request structure across model types, reducing integration work when switching or adding models. - The playground mirrors API fields, which shortens the path from experimentation to production code. - Usage-based billing against a single balance simplifies cost tracking across multiple models. Cons: - Webhook delivery is best-effort; final status and result URLs must be confirmed via the tasks endpoint. - Seedance 2.5 text-to-video durations are limited to 4 to 30 seconds with resolutions of 480p, 480p-draft, 720p, or 1080p. - The 1080p upscale via draft requires submitting the preview task_id within 7 days of creation. 【FAQ】 Q: How do I check the status of a long-running generation task? A: Poll the tasks endpoint by task ID. Status values are 0 for running, 1 for succeeded, and 2 for failed. Q: What happens when a task reaches a terminal state? A: SJolt sends a POST to your top-level webhook URL. Because delivery is best-effort, query the tasks endpoint for final status and result URLs. Q: What are the duration and resolution options for Seedance 2.5 text-to-video? A: Videos can be 4 to 30 seconds long at resolutions of 480p, 480p-draft, 720p, or 1080p. A 480p-draft preview can be upscaled to 1080p by submitting the preview task_id within 7 days of creation.
usage
WPS AI is an AI assistant built into WPS Office that helps users write documents, generate presentation outlines and…
Baidu Comate is an AI coding assistant built by Baidu on its ERNIE model. It goes beyond simple autocomplete: it…
UniConverter is an AI media conversion and enhancement tool for video, audio, and images. It combines everyday…
6 similar tools to consider instead of SJolt