ByteDance / seed-v2-pro-uncensored
seed-v2-pro-uncensored
ByteDance Seed V2 Pro Uncensored is a long-context multimodal chat model for complex reasoning and image Q&A. It generates detailed text answers from prompts and images.
Overview
Seed V2 Pro Uncensored is a text model from ByteDance’s Seed family. It’s designed for complex, long-horizon tasks and agent-style workflows. (seed.bytedance.com)
It accepts text and can also read images for visual reasoning. It produces text responses that can include step-by-step plans, structured extractions, and detailed explanations. This helps when you need one model for both analysis and visual Q&A. (seed.bytedance.com)
What you can build
- Screenshot Q&A for apps, dashboards, and product UIs (seed.bytedance.com)
- Chart and table interpretation with plain-language summaries (seed.bytedance.com)
- Document-style extraction from images into clean bullet points (seed.bytedance.com)
- Multi-step troubleshooting plans and decision trees (github.com)
- Coding support for long tasks, like repo-level change planning (github.com)
- “Plan → act → reflect” style task decomposition for research and synthesis (github.com)
Inputs
- A required user request in plain text. Put the full task, constraints, and desired format in one message.
- Optional reference images for visual understanding. You can attach multiple images, separated with semicolons, up to 20 images.
- Optional user identifier as a number or string. Use it to keep chat history tied to one user.
- Optional session identifier as a number or string. Use it to group a single multi-turn session.
- Optional system-level instructions in plain text. Use this to set role, tone, and hard rules.
- Optional creativity control. Lower values make outputs steadier. Higher values add variety.
- Optional nucleus sampling control. Lower values narrow word choice. Higher values broaden it.
- Optional repetition controls. You can penalize repeated wording and push topic novelty.
- Optional maximum response length in tokens. Very large outputs can hit platform or provider limits.
- Optional reasoning depth selection. Choose from minimal, low, medium, or high.
- Optional deep thinking mode. Choose enabled, disabled, or auto.
Outputs
The model returns a text response.
The response is typically a single assistant message. It can include: - A direct answer
- Lists and checklists
- Structured text meant to be parsed as JSON if you request it
- Explanations tied to details it finds in attached images (seed.bytedance.com)
Limitations
- This is a text model. It does not generate images, audio, or video.
- The Seed 2.0 Pro-class deployment is documented with a 256K context window. Requests that exceed limits can truncate or fail. (docs.byteplus.com)
- Documented maximum output (including reasoning content) can be up to 128K tokens in some providers. Your Wiro settings may cap outputs lower. (docs.byteplus.com)
- Visual understanding depends on image quality. Blurry screenshots, heavy compression, or skewed scans can cause missed details.
- Multi-image prompts can reduce accuracy when images conflict or lack clear ordering.
- Like other LLMs, it can hallucinate. Verify factual claims, math, and citations.
Safety & compliance
Even if this listing is labeled “uncensored,” you must follow the BytePlus GenAI Acceptable Use Policy when the underlying service is BytePlus-hosted.
Key restrictions called out by BytePlus include: - Don’t use it to support illegal activity or regulated wrongdoing.
- Don’t create or enable sexual content involving minors.
- Don’t violate privacy or process personal data without a legal basis.
- Don’t identify people or verify identity from biometric or personal data.
- Don’t generate impersonation or likeness content without consent.
- Don’t run scams, phishing, or deceptive misinformation.
- Don’t remove or hide model watermarks or provenance signals.
- Don’t bypass or try to defeat safety mechanisms. (docs.byteplus.com)
BytePlus also expects responsible deployment. Test your workflows, keep human review, and disclose AI-generated content where required. (docs.byteplus.com)
API quick start
Run seed-v2-pro-uncensored with a single API call.
{
"prompt": "Tell me about Türkiye",
"inputImage": "https://your-cdn.com/input.png",
"userId": "...",
"sessionId": "..."
}