API Reference

Drop-in OpenAI compatible. Same SDK. Same format. Change the base URL and access 10+ models from every major provider.

Quick Start

Install the OpenAI SDK (or any OpenAI-compatible client) and change the base URL to https://ai.empire325marketing.com/v1. That's it.

# Install the OpenAI SDK pip install openai # 325 API — identical interface, every model from openai import OpenAI client = OpenAI( api_key="hm-...", base_url="https://ai.empire325marketing.com/v1" ) response = client.chat.completions.create( model="325-auto", messages=[{"role": "user", "content": "Explain quantum computing"}] ) print(response.choices[0].message.content)

Also works with any OpenAI-compatible client: LangChain, LlamaIndex, Vercel AI SDK, etc.

Authentication

All requests require an API key passed in the Authorization header. Keys use the hm- prefix.

curl https://ai.empire325marketing.com/v1/chat/completions \ -H "Content-Type: application/json" \ -H "Authorization: Bearer hm-your-api-key" \ -d '{ "model": "325-balanced", "messages": [{"role": "user", "content": "Hello"}] }'

Get your API key by signing up for a plan. Keys are rate-limited per your subscription tier.

Model Tiers

325 API provides four model tiers, each optimized for different workloads:

325-fast

Sub-second responses via Cerebras + Groq. Best for HTML, CSS, Shell, SQL, simple queries.

325-balanced

DeepSeek V4 Pro with Claude fallback. Best for Python, code, math, reasoning.

325-ultra

Cascade ensemble. Multi-model consensus. Best for analysis, research, OSINT.

325-auto — Smart domain routing. Analyzes your prompt and picks the optimal tier automatically. Saves 91% on tokens vs always using the most expensive model.

Endpoints

POST	`/v1/chat/completions`	Chat completions — OpenAI-compatible
POST	`/v1/research`	Multi-source research with citations (Pro/Max)
POST	`/v1/code/build`	Code generation via Claude Opus 4.8 (Pro/Max)
GET	`/v1/models`	List available models for your plan
GET	`/health`	API health check
GET	`/v1/dashboard`	Your usage dashboard (authenticated)

/v1/chat/completions

Standard OpenAI-compatible chat completions endpoint. All parameters match the OpenAI spec.

model	Required. One of: `325-fast`, `325-balanced`, `325-ultra`, `325-auto`
messages	Required. Array of message objects with `role` and `content`
max_tokens	Maximum tokens in response. Default: 2000
temperature	Sampling temperature 0-2. Default: 0.7
stream	Enable streaming (coming soon). Default: false

# Python example response = client.chat.completions.create( model="325-balanced", messages=[ {"role": "system", "content": "You are a helpful assistant."}, {"role": "user", "content": "Write a Python function to reverse a linked list."} ], max_tokens=1000, temperature=0.3 )

/v1/research

Multi-source web research with synthesized answers and source citations. Available on Pro and Max plans.

query	Required. Research question
depth	`standard` (5 sources) or `deep` (8 sources). Default: standard

curl https://ai.empire325marketing.com/v1/research \ -H "Authorization: Bearer hm-..." \ -H "Content-Type: application/json" \ -d '{"query": "Latest advances in AI safety research 2026", "depth": "deep"}'

Rate Limits

Limits are per API key, per minute. Exceeded limits return HTTP 429 with a Retry-After header.

Starter	30 requests/minute
Pro	120 requests/minute
Max	500 requests/minute

Token quotas reset on your monthly billing date. We notify at 80% and 95% usage. No overage charges — requests return 429 until reset or upgrade.

Errors

401	Invalid or missing API key
429	Rate limit or quota exceeded. Check Retry-After header
400	Invalid request body or parameters
503	Service temporarily unavailable (provider outage — auto-failover typically handles this)