BYOK & ZERO-MARKUP GUIDES
Connect your favorite AI Platform.
Crescelo guarantees 0% token markup. You can bring your own API key and pay raw wholesale rates directly to providers, or use our built-in Zero-Cost failover engine for $0 compute.
STEP-BY-STEP GUIDES
Select an AI model connector.
Each guide walks you through account creation, API key generation, budget configuration, and Crescelo dashboard connection.
Zero-Cost Free Engine
Intelligent multi-provider cascade across Groq Free Tier (Llama 3.3 70B), Google Gemini Flash, and local Ollama. Zero credit card or subscription needed.
OpenAI (GPT-4o & GPT-4o-mini)
Direct connection to platform.openai.com. Pay raw wholesale rates directly to OpenAI with zero token markup and enterprise AES-256 key encryption.
Anthropic Claude (3.5 Sonnet & Haiku)
Connect via Anthropic Console. Renowned for natural human-like voice, low hallucination rates, and superior brand safety guardrails.
Google Gemini (Flash & Pro)
Get a free API key from Google AI Studio. Gemini Flash processes huge document catalogs and complex price sheets without truncation.
Groq Cloud (Llama 3.3 70B)
Runs on custom LPU tensor architecture. Streams complete multi-paragraph answers before standard models finish their first word.
DeepSeek (V3 & R1 Reasoning)
Frontier model intelligence at a fraction of standard cost. Supports deep reasoning mode (R1) for complex multi-variable quoting.
Local Server (Ollama Offline)
Run Llama 3.2, Mistral, or Qwen on your own computer or private cloud instance. Zero external network calls; 100% private inference.
Need assistance setting up a custom endpoint?
We support custom OpenAI-compatible enterprise endpoints (Azure OpenAI, AWS Bedrock proxies, and private vLLM clusters).
Contact Engineering Team