Fable-Fusion-711 — API endpoint

OpenAI-compatibleno API keystreaming

Base URL

https://investigation-approved-residence-cruz.trycloudflare.com/v1

checking status…

Not a stable endpoint. This runs on a single personal GPU behind a Cloudflare quick tunnel. The URL above changes whenever the tunnel restarts, and it can go offline without notice — this is a hobby box, not a production service. Don't build anything that depends on it staying up.

Model

/root/models/Qwen3.6-27B-Fable-Fus-711-UnHeretic-NM-DAU-NEO-MAX-NEO-MTP-Q4_K_M.gguf

Qwen3.6-27B, Q4_K_M quant, MTP self-speculative decoding, served by llama.cpp. Reasoning model — pass chat_template_kwargs: {"enable_thinking": false} to skip the thinking trace.

curl

curl https://investigation-approved-residence-cruz.trycloudflare.com/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "/root/models/Qwen3.6-27B-Fable-Fus-711-UnHeretic-NM-DAU-NEO-MAX-NEO-MTP-Q4_K_M.gguf",
    "messages": [{"role": "user", "content": "Say hello in five words."}],
    "temperature": 0.7,
    "max_tokens": 256,
    "chat_template_kwargs": {"enable_thinking": false}
  }'

Python (OpenAI SDK)

from openai import OpenAI

client = OpenAI(
    base_url="https://investigation-approved-residence-cruz.trycloudflare.com/v1",
    api_key="not-needed",
)

resp = client.chat.completions.create(
    model="/root/models/Qwen3.6-27B-Fable-Fus-711-UnHeretic-NM-DAU-NEO-MAX-NEO-MTP-Q4_K_M.gguf",
    messages=[{"role": "user", "content": "Say hello in five words."}],
    temperature=0.7,
    max_tokens=256,
    extra_body={"chat_template_kwargs": {"enable_thinking": False}},
)
print(resp.choices[0].message.content)

JavaScript (fetch, streaming)

const res = await fetch("https://investigation-approved-residence-cruz.trycloudflare.com/v1/chat/completions", {
  method: "POST",
  headers: { "Content-Type": "application/json" },
  body: JSON.stringify({
    model: "/root/models/Qwen3.6-27B-Fable-Fus-711-UnHeretic-NM-DAU-NEO-MAX-NEO-MTP-Q4_K_M.gguf",
    messages: [{ role: "user", content: "Say hello in five words." }],
    stream: true,
    chat_template_kwargs: { enable_thinking: false }
  })
});
// res.body is an SSE stream of `data: {...}` chunks, OpenAI-style

Limits

Context window	8192 tokens
Concurrency	4 slots (shared GPU — expect queueing under load)
Auth	none
Uptime	best-effort, no guarantee

Endpoints follow the standard OpenAI Chat Completions API — /v1/chat/completions, /v1/models, /v1/completions. Full parameter list in the llama.cpp server docs.