API Console

Darkbloom API — OpenAI-compatible. Swap your base URL, keep your existing code. Every request is end-to-end encrypted and processed on hardware-attested Apple Silicon.

Endpoint Reference

Expand each endpoint to see request/response format and notes.

Base URL

https://api.darkbloom.dev/v1

All endpoints are relative to this base URL. Provider attestation and pricing endpoints are publicly accessible without authentication.

API Key

No API key generated

Use this key in the Authorization: Bearer header for all authenticated requests.

Quick Start

Install the OpenAI SDK or Vercel AI SDK. The Darkbloom API is fully OpenAI-compatible — just change the base URL.

# No installation needed
export DARKBLOOM_API_KEY="<YOUR_API_KEY>"
export DARKBLOOM_BASE_URL="https://api.darkbloom.dev/v1"

Available Models

Model ID	Type	Architecture
mlx-community/gemma-4-26b-a4b-it-8bit	text	26B MoE, 4B active — recommended
qwen3.5-27b-claude-opus-8bit	text	27B dense, Claude Opus distilled
mlx-community/Qwen3.5-122B-A10B-8bit	text	122B MoE, 10B active
mlx-community/MiniMax-M2.5-8bit	text	239B MoE, 11B active
CohereLabs/cohere-transcribe-03-2026	stt	2B conformer

Model availability depends on online providers. Check /v1/models for real-time availability.

Chat Completions

Stream chat completions with any supported model. Supports system messages, multi-turn conversations, and thinking/reasoning output.

curl -X POST https://api.darkbloom.dev/v1/chat/completions \
  -H "Authorization: Bearer <YOUR_API_KEY>" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "mlx-community/gemma-4-26b-a4b-it-8bit",
    "messages": [{"role": "user", "content": "Explain quantum computing"}],
    "stream": true,
    "max_tokens": 1024
  }'

Image Generation

Generate images with FLUX models running on Metal-accelerated Apple Silicon. Returns base64-encoded PNG.

import base64
from openai import OpenAI

client = OpenAI(base_url="https://api.darkbloom.dev/v1", api_key="<YOUR_API_KEY>")

response = client.images.generate(
    model="flux_2_klein_4b_q8p.ckpt",
    prompt="A serene mountain landscape at sunset",
    n=1,
    size="1024x1024",
)

# Save the image
img_data = base64.b64decode(response.data[0].b64_json)
with open("output.png", "wb") as f:
    f.write(img_data)

Speech-to-Text

Transcribe audio files using the Cohere Transcribe model. Supports WAV, MP3, WebM, M4A, and FLAC.

from openai import OpenAI

client = OpenAI(base_url="https://api.darkbloom.dev/v1", api_key="<YOUR_API_KEY>")

with open("audio.wav", "rb") as f:
    transcript = client.audio.transcriptions.create(
        model="CohereLabs/cohere-transcribe-03-2026",
        file=f,
    )

print(transcript.text)

List Models

Check available models, provider counts, and attestation status.

from openai import OpenAI

client = OpenAI(base_url="https://api.darkbloom.dev/v1", api_key="<YOUR_API_KEY>")

models = client.models.list()
for model in models.data:
    print(f"{model.id}")