code.ts
import { OpenAI } from "openai"

const client = new OpenAI({
  baseURL: "https://core.zaguan.ai/v1",
  apiKey: process.env.ZAGUAN_KEY
})

// Use any provider - same interface
const response = await client.chat.completions.create({
  model: "anthropic/claude-3-sonnet",
  messages: [{ role: "user", content: "Write production code" }]
})

// Switch providers in one line - no refactoring
const mistralResponse = await client.chat.completions.create({
  model: "mistral/mistral-large",
  messages: [{ role: "user", content: "Explain this code" }]
})

The Translation Layer That Just Works

Other gateways forward your errors. Zaguán actively translates your OpenAI-style calls to work optimally with each provider's specific API, parameters, and features. One integration, 500+ models.

Expert AI Personas for Professional Work

Instead of getting generic responses from a single AI model, Orchestrator delivers specialized, high-quality outputs through expert AI personas—each designed for specific professional tasks.

✍️

Wordsmith

Marketing Copywriter

Persuasive, conversion-focused marketing copy with deep expertise in consumer psychology. Perfect for ad copy, landing pages, and campaigns.

🏗️

Architect

Software Engineer

Production-ready code with clean architecture and comprehensive documentation. Acts like a principal engineer who thinks through the entire solution.

🧠

Sage

Strategic Advisor

Wise, authoritative strategic advice. The tone of a seasoned board member or mentor. Ideal for strategic playbooks and leadership decisions.

Standard, Fast & Ultra Personas

Choose the right balance of quality, speed, and cost for your use case.

Standard Personas

Premium quality for professional work

Best for: Professional deliverables, high-quality output

Speed: 30-60 seconds

Cost: 1x credits

Fast Personas

Optimized for speed and efficiency

Best for: Iteration, prototyping, quick drafts

Speed: 15-30 seconds

Cost: 0.5x credits

Ultra Personas

Mission-critical with zero-error tolerance

Best for: Legal, compliance, finance, architecture

Speed: 120-300 seconds

Cost: 5x credits

How the Five-Step Workflow Works

Each persona runs a multi-step reasoning loop inspired by how expert teams collaborate.

1

Understanding

Analyzes your request and determines the best approach

2

Planning

Creates a strategic blueprint for the response

3

Execution

Generates a comprehensive first draft

4

Review

Critically evaluates and identifies improvements

5

Refinement

Produces the final, polished output

Two Products, One Platform

Zaguán is two things: an expert reasoning layer and a multi-provider gateway. You can use them together or separately.

🎯

Orchestrator

Reasoning-as-a-Service (RaaS)

Expert AI personas for professional work. Get specialized, high-quality outputs through multi-step workflows designed for specific tasks.

Perfect for:

  • • Marketing copy and content creation
  • • Software development and architecture
  • • Strategic planning and analysis
  • • Legal, compliance, and financial work
Learn About Orchestrator
🌉

Zaguán Core

Multi-Provider AI Gateway

OpenAI-compatible gateway to 500+ models from 15+ providers. Smart translation layer adapts your requests for optimal results with each provider.

Perfect for:

  • • Building with multiple AI providers
  • • Avoiding vendor lock-in
  • • Testing different models easily
  • • Automatic failover and routing
Learn About Core Gateway

Use them together: Build your app with Zaguán Core for flexibility, and use Orchestrator personas for specialized tasks that need expert-level quality.

One API. Every Model. Zero Glue Code.

A zaguán is an architectural passage between the street and the home. We are the intelligent passage between your code and every AI model on earth.

The Translation Layer is the Gold

Other gateways just forward your requests. Zaguán actively translates your OpenAI-style calls to work optimally with each provider's specific API, parameters, and features.

OpenAI-Compatible

Swap your base URL and start using multiple providers in minutes. No code refactoring required.

Smart Translation

We normalize and enrich requests per model: system prompts, parameters, safety options.

No Vendor Lock-In

Switch providers in seconds. Automatic failover means zero downtime when providers fail.

Supported Providers

Access 500+ models from leading AI providers

OpenAI
Anthropic
Google
Mistral
Alibaba
Inception
View All Supported Models

What You Get

Same across both products — enterprise-grade infrastructure and seamless integration.

It's Just OpenAI SDK

# Same code works with 15+ providers
from openai import OpenAI

client = OpenAI(
  base_url="https://core.zaguan.ai/v1",
  api_key="your-zaguan-key"
)

# Use any model
client.chat.completions.create(
  model="anthropic/claude-3-sonnet",
  messages=[{"role": "user", "content": "Hello"}]
)

Predictable Outputs

Multi-step workflow ensures consistent, professional-grade outputs. No more random AI responses.

Any Language

# Responds in your language automatically
response = client.chat.completions.create(
  model="zaguan/standard",
  messages=[
    {"role": "user", "content": "Escribe un poema"}
  ]
)
# Returns Spanish response without prompting

Real-Time Streaming

# Watch AI think in real-time
stream = client.chat.completions.create(
  model="zaguan/standard",
  messages=[...],
  stream=True
)

for chunk in stream:
  print(chunk.choices[0].delta.content)

Built by a Focused Specialist, Not a VC Startup.

No VC pressure, no need to chase growth metrics, no risk of enshittification. Just a bootstrapped business that makes money by being useful.

Predictable & Simple

# One flat price, no surprises
# €15/month = 1M credits (Founder's Plan)
# Use credits across any provider/model
# No per-token charges, no hidden fees

One flat monthly price. No token billing surprises, no hidden costs. You know exactly what you'll pay.

No Setup Headaches

# Zero configuration - just one key
import { OpenAI } from "openai"

const client = new OpenAI({
  baseURL: "https://core.zaguan.ai/v1",
  apiKey: process.env.ZAGUAN_KEY  // Only key you need
})

Just sign up and start building. We handle provider keys, rate limits, and API versions — you write code.

Stable & Long-Term

profitable since 2024
bootstrapped, no VC pressure
founder-led, not a side project

Self-sustaining infrastructure that won't disappear. Built for the long haul.

Provider-Native Support

# Full access to provider features
# Anthropic: messages API, structured outputs
# Google: system instructions, tuning
# Mistral: custom models, fine-tuning
# All through the same OpenAI-compatible interface

Full access to each provider's features through a unified interface.

Production-Ready SDKs

Official SDKs for Python, TypeScript, and Go. Type-safe, fully documented, and battle-tested with comprehensive test coverage.

🐍pip install zaguan-sdk
📘npm i @zaguan_ai/sdk
🔷go get zaguan-sdk-go
View SDK Documentation
👤

Direct Founder Access & Real Support

Clear SDKs, real examples in popular stacks (Node, Python, Go), and responsive support from the creator. Early users help shape the roadmap — you're not just a customer.

Straightforward Pricing

From side projects to enterprise scale - access 500+ AI models through one elegant gateway. No token counting, no surprise bills.

Side Project

Perfect for hobbyists and experimenters

$39/month
Monthly Allowance
400 Credits
  • 400 credits/month
  • All AI models
  • API access
  • Email support
Most Popular

Startup

For MVPs and early-stage products

$99/month
Monthly Allowance
1000 Credits
  • 1,000 credits/month
  • All AI models
  • API access
  • Priority support
  • Usage analytics

Scale

For teams and agencies

$249/month
Monthly Allowance
2500 Credits
  • 2,500 credits/month
  • All AI models
  • API access
  • Priority support
  • Usage analytics
  • Team management

Pro Tip: Need more credits? Top-up packs are available in your dashboard and never expire.

One API Key, 500+ Models

No more juggling provider dashboards. Get a single endpoint, one bill, and direct access to the founder.

Frequently Asked

Enterprise features?

Today I'm focused on solo devs, small teams, and agencies. Enterprise features (custom contracts, SSO, on-prem) come when you ask for them. Email me if you need them — real demand shapes the roadmap.

How is Orchestrator different from ChatGPT or Claude?

ChatGPT and Claude are general-purpose AI assistants. Orchestrator provides specialized expert personas with predictable quality through a multi-step workflow. Instead of one model doing everything, each persona uses multiple models in sequence (Understanding → Planning → Execution → Review → Refinement) to deliver consistent, professional-grade outputs optimized for specific tasks.

What's on the roadmap?

What I'm building next:

  • More specialized personas (request ones you need)
  • Real-time streaming and collaboration
  • Custom workflow builder for power users
  • Shared persona libraries for teams

As a Founder member, you'll be first to test new personas and help shape what we build next.

No Newsletters, Just Product Updates

New models, provider changes, and feature releases — nothing else. No marketing spam, no "thought leadership."