XField is an API for uncensored open-weight language models. The models answer the question you asked, without added moderation layers or content filters in between. It is built for security researchers, red teams, trust and safety teams, and anyone running coding agents who is tired of refusals getting in the way of real work.
It speaks both the OpenAI and the Anthropic APIs. If you already use the OpenAI SDK, the Anthropic SDK, Claude Code, Codex, OpenCode, or any other compatible client, you change two things: the base URL and the model id. Everything else, from streaming to tool calls, works the way you expect.
Start here
QuickstartGet a key and send your first request in a few minutes.ModelsThe four models, their limits, and which one to pick.Claude CodePoint Claude Code at XField with two environment variables.Limits and errorsStatus codes, rate limits, and when to retry.Guides
Security researchExploit analysis, detection rules, and reports without the hedging.Red teamingPlan engagements and generate adversarial test cases.Trust and safetyClassify, label, and study harmful content at scale.At a glance
| Value | |
|---|---|
| OpenAI base URL | https://xfield.cloud.brutal.ai/api/v1 |
| Anthropic base URL | https://xfield.cloud.brutal.ai/api |
| Auth | Authorization: Bearer <key> or x-api-key: <key> |
| Protocols | Chat Completions, Responses, Messages, legacy Completions |
| Models | DeepSeek V4.1 Flash, Kimi K3.1 Pro, GLM 5.3, GLM 5.3 Flash, Qwen 3.8 27B |
| Streaming | Server-sent events on every route |
| Data | Prompts and completions are not stored or used for training |
Access is by invitation. If you do not have an account yet, request access.