Skip to main content
A unified, compliance-ready gateway for the world’s best models. Integrate Claude, GPT-4, and open-source models with built-in PII redaction, rate limiting, and audit trails.

​
Why use the Gateway?

You could call Anthropic or OpenAI directly. But then you have to build the compliance layer yourself.
  • Zero-Retention Agreements: We hold BAAs with major providers—pass this through to your users.
  • Unified Audit Log: Every prompt and completion logged for compliance (optional).
  • PII Redaction: Automatically strip names, SSNs, and dates before they hit the model.
  • Provider Fallback: If OpenAI is down, we route to Azure or Anthropic automatically.

​
Quick example

import Casedev from 'casedev';

const client = new Casedev({ apiKey: process.env.CASEDEV_API_KEY });

// Process a request from your user
const response = await client.llm.v1.chat.createCompletion({
  model: 'anthropic/claude-sonnet-4.5',
  messages: [
    { role: 'user', content: userPrompt }
  ]
});

// Return the response to your user
console.log(response.choices[0].message.content);

​
Specialized Models

Choose the right model for your feature:
ProviderModelsBest for
AnthropicClaude 3.5 Sonnet, OpusContract Review. Largest context window, best reasoning.
OpenAIGPT-4oExtraction. Reliable JSON output for structured data.
GoogleGemini 1.5 ProLarge Documents. 1M+ token context for analyzing entire case files.
CaseMarkLegal-BERT / RoBERTaClassification. Fine-tuned for specific legal taxonomy.
See Models for the full list.

​
Endpoints

Models

GET /llm/v1/models — List available models

Chat completions

POST /llm/v1/chat/completions — Generate AI responses

Embeddings

POST /llm/v1/embeddings — Generate vector embeddings

Vaults

Combine LLMs with semantic search for RAG-powered document Q&A

Web Search

Ground LLM responses with real-time web research and citations

Workflows

Chain LLM calls with other services in automated pipelines

Format

Convert LLM outputs into polished PDFs and Word documents