⚡ PixelRouter v1.4.0 — High-Speed OpenAI & Anthropic Claude AEO Gateway

🎁 Claim Free Production Requests & Live Dashboard Key:
👉 Get Free Production API Key & Track Real-Time Telemetry at pixeloffice.eu/dashboard.html

PixelRouter is a high-speed, enterprise-hardened AI gateway and autonomous agent infrastructure layer powered by Pixel Office EU and the BLUN Intelligence Engine. It provides native 1-line drop-in compatibility for both OpenAI REST and Anthropic Claude /v1/messages, sub-35ms edge latency, exact-match prompt caching (1ms replays), and an Autonomous Agent Highway.

Gateway Base URL: https://api.pixeloffice.eu/v1
Anthropic Endpoint: https://api.pixeloffice.eu/v1/messages
Machine Web Highway: https://api.pixeloffice.eu/v1/reader
NPM Package: @pixeloffice-eu/router (v1.4.0)
PyPI Package: pixeloffice-router (v1.4.0)
VS Code & Cursor Extension: pixelrouter-vscode (v1.4.0)
Live Interactive Playground: https://pixeloffice.eu/router.html
Status: Live & Production Ready (HTTP 200 OK, Sub-35ms Edge)

🌟 Major Features in v1.4.0

1. Native Anthropic Claude `/v1/messages` Compatibility

Directly compatible with official Anthropic SDKs (@anthropic-ai/sdk, anthropic Python, Claude Code, Cursor). Point the base URL to PixelRouter and authenticate using x-api-key:

import anthropic

client = anthropic.Anthropic(
    base_url="https://api.pixeloffice.eu",
    api_key="px_test_free" # Or your Pixel Office production key
)

message = client.messages.create(
    model="claude-3-haiku-20240307",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Explain quantum teleportation in 2 sentences."}]
)
print(message.content[0].text)

2. Autonomous Agent Web Highway (`POST /v1/reader`)

Autonomous agents break in production not from lack of intelligence, but from bloated HTML, cookies, and tracking scripts. The /v1/reader endpoint strips noisy DOM nodes and returns clean Markdown with Top 1KB Fact Anchors, reducing token consumption by up to 98%:

curl -X POST https://api.pixeloffice.eu/v1/reader \
  -H "Content-Type: application/json" \
  -d '{"url":"https://example.com"}'

3. Zero-Drop Handshake Failover & Loop Guardrails

15ms Sentinel Failover: When an upstream model provider returns HTTP 429 or 5xx outage, PixelRouter transparently re-routes to Google Gemini 2.5 Flash within 15ms so long-running agent workflows never abort.
Loop & Budget Guardrail: Canonicalizes JSON tool call arguments into key-sorted hashes. If 3 consecutive identical tool calls are detected, execution halts to protect your budget from runaway loops.

4. 1-Line LiteLLM Proxy Integration

Drop PixelRouter into your existing LiteLLM proxy in 1 line of YAML:

model_list:
  - model_name: gpt-4o
    litellm_params:
      model: openai/blun-auto
      api_base: https://api.pixeloffice.eu/v1
      api_key: os.environ/PIXELROUTER_API_KEY

🚀 Quickstart Integrations

Python (OpenAI SDK / LangChain / CrewAI)

pip install pixeloffice-router openai

from openai import OpenAI

client = OpenAI(
    base_url="https://api.pixeloffice.eu/v1",
    api_key="px_test_free"
)

response = client.chat.completions.create(
    model="blun-auto", # Auto-routes to DeepSeek V4 Flash, GLM-5.3, or Gemini
    messages=[{"role": "user", "content": "Write a high-performance TypeScript debounce utility."}],
    stream=True
)

for chunk in response:
    if chunk.choices and chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="", flush=True)

1-Click Cursor IDE & VS Code Setup

npx @pixeloffice-eu/router

Or set manually in Cursor Settings > Models > OpenAI API Key & Base URL:

Base URL: https://api.pixeloffice.eu/v1
Model: blun-auto (or deepseek-v4-flash, claude-3.5-sonnet)
API Key: px_test_free

📊 Model Fleet & Pricing Reference

Model Identifier	Engine Architecture	Latency	Cache Replay	Context Window
`blun-auto`	Autonomous SmartRouter	< 35 ms	1 ms	1,000,000
`deepseek-v4-flash`	DeepSeek 284B MoE	< 25 ms	1 ms	1,000,000
`glm-5.3-flash`	Zhipu AI High-Speed Reasoning	< 25 ms	1 ms	128,000
`claude-fable-5.1`	Anthropic Ultra-Frontier	~ 450 ms	1 ms	1,000,000
`claude-3.5-sonnet`	Anthropic Coding SOTA	~ 550 ms	1 ms	200,000
`deepseek-v4-pro`	DeepSeek 1.6T MoE Deep Reasoning	~ 650 ms	1 ms	1,000,000
`google/gemini-2.5-flash`	Google Gemini Multimodal	< 40 ms	1 ms	1,000,000

🛡️ License & Enterprise Verification

Maintained by Pixel Office EU. Licensed under the MIT License.

Downloads last month: -; Downloads are not tracked for this model. How to track