Getting Started

Integrate RyanChat's powerful multi-LLM API into your application in minutes.

Quick Start

1. Get Your API Key

2. Make Your First Request

Use your API key to make requests to the chat completions endpoint.

bash

curl https://dev.ryanchat.com/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.1",
    "messages": [
      {
        "role": "user",
        "content": "Hello! How are you?"
      }
    ]
  }'

# For vision-enabled models with uploaded images:
curl https://dev.ryanchat.com/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.1",
    "messages": [
      {
        "role": "user",
        "content": [
          {
            "type": "text",
            "text": "What's in this image?"
          },
          {
            "type": "image_url",
            "image_url": {
              "url": "https://dev.ryanchat.com/v1/files/550e8400-e29b-41d4-a716-446655440000.jpg"
            }
          }
        ]
      }
    ]
  }'

3. Handle the Response

The API returns a standardized response format compatible with OpenAI's API.

json

{
  "id": "chatcmpl-1234567890",
  "object": "chat.completion",
  "created": 1677652288,
  "model": "gpt-5.1",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "Hello! I'm doing well, thank you for asking. How can I help you today?"
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 13,
    "completion_tokens": 20,
    "total_tokens": 33
  }
}

API Endpoints

POST/v1/chat/completions

Chat Completions

Generate chat completions using any supported model.

Parameter	Type	Required	Description
`model`	`string`	Required	The model to use for completion (e.g., "gpt-5.1", "claude-sonnet-4-5", "gemini-2.5-pro")
`messages`	`array`	Required	A list of messages comprising the conversation. For vision-enabled models, include image URLs from file uploads.
`max_tokens`	`integer`	Optional	Maximum number of tokens to generate
`temperature`	`number`	Optional	Controls randomness in the response (0.0 to 2.0)
`response_format`	`object`	Optional	Specify the response format for structured output

Example Response

json

{
  "id": "chatcmpl-1234567890",
  "object": "chat.completion",
  "created": 1677652288,
  "model": "gpt-5.1",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "Hello! I'm doing well, thank you for asking. How can I help you today?"
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 13,
    "completion_tokens": 20,
    "total_tokens": 33
  }
}

Using Files with Vision Models

Vision-enabled models like GPT-5.1, Claude Sonnet 3.5, and Gemini Pro Vision can analyze uploaded images. Use the file upload endpoint to upload images, then reference them in your messages using the returned file URL. Some models also support document analysis for PDFs and other file types.

json

{
  "model": "gpt-5.1",
  "messages": [
    {
      "role": "user",
      "content": [
        {
          "type": "text",
          "text": "What's in this image?"
        },
        {
          "type": "image_url",
          "image_url": {
            "url": "https://dev.ryanchat.com/v1/files/550e8400-e29b-41d4-a716-446655440000.jpg"
          }
        }
      ]
    }
  ]
}

POST/v1/files/upload

Upload Files

Upload files for AI processing. Files can be used with vision-enabled models or document analysis.

Parameter	Type	Required	Description
`files`	`file[]`	Required	One or more files to upload. Supported formats: Images (JPEG, PNG, GIF, WebP), Documents (PDF, DOC, DOCX, XLS, XLSX, TXT), and other common file types.

Supported File Types

Images: JPEG, PNG, GIF, WebP, BMP, TIFF

Documents: PDF, DOC, DOCX, XLS, XLSX, PPT, PPTX

Text: TXT, CSV, JSON, Markdown, HTML

Size Limit: 50MB per file

Example Request

bash

curl -X POST https://dev.ryanchat.com/v1/files/upload \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -F "files=@document.pdf" \
  -F "files=@image.jpg"

Example Response

json

{
  "success": true,
  "files": [
    {
      "id": "550e8400-e29b-41d4-a716-446655440000",
      "filename": "550e8400-e29b-41d4-a716-446655440000.pdf",
      "originalName": "document.pdf",
      "mimetype": "application/pdf",
      "size": 245760,
      "url": "/api/v1/files/550e8400-e29b-41d4-a716-446655440000.pdf",
      "uploadDate": "2025-11-27T04:30:00.000Z"
    },
    {
      "id": "550e8400-e29b-41d4-a716-446655440001",
      "filename": "550e8400-e29b-41d4-a716-446655440001.jpg",
      "originalName": "image.jpg",
      "mimetype": "image/jpeg",
      "size": 512000,
      "url": "/api/v1/files/550e8400-e29b-41d4-a716-446655440001.jpg",
      "uploadDate": "2025-11-27T04:30:00.000Z"
    }
  ],
  "message": "Successfully uploaded 2 file(s)"
}

GET/v1/files/:filename

Serve File

Retrieve an uploaded file by its filename.

Parameter	Type	Required	Description
`filename`	`string`	Required	The filename returned from the upload endpoint (UUID-based filename).

Example Request

bash

curl https://dev.ryanchat.com/v1/files/550e8400-e29b-41d4-a716-446655440000.pdf \
  -H "Authorization: Bearer YOUR_API_KEY"

Response

Returns the file content with appropriate Content-Type headers. Files are cached for 1 year.

POST/v1/chat/completions (Structured Output)

Chat Completions with Structured Output

Generate chat completions with guaranteed JSON response format using JSON Schema.

Parameter	Type	Required	Description
`model`	`string`	Required	The model to use for completion
`messages`	`array`	Required	A list of messages comprising the conversation
`response_format`	`object`	Required	Defines the structure of the response

Structured Output Example

Use JSON Schema to define the exact structure you need for the response. The AI will strictly follow the schema you provide.

json

{
  "model": "gpt-5.1",
  "messages": [
    {
      "role": "user",
      "content": "Extract information about a person from this text: John Smith is 35 years old and works as a software engineer in San Francisco."
    }
  ],
  "response_format": {
    "type": "json_schema",
    "json_schema": {
      "name": "person_info",
      "schema": {
        "type": "object",
        "properties": {
          "name": {
            "type": "string",
            "description": "The person's full name"
          },
          "age": {
            "type": "integer",
            "description": "The person's age in years"
          },
          "occupation": {
            "type": "string",
            "description": "The person's job title"
          },
          "location": {
            "type": "string",
            "description": "The city where the person lives"
          }
        },
        "required": ["name", "age", "occupation", "location"],
        "additionalProperties": false
      },
      "strict": true
    }
  }
}

Guaranteed Response Format

json

{
  "id": "chatcmpl-structured-1234567890",
  "object": "chat.completion",
  "created": 1677652288,
  "model": "gpt-5.1",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "{\"name\": \"John Smith\", \"age\": 35, \"occupation\": \"software engineer\", \"location\": \"San Francisco\"}"
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 45,
    "completion_tokens": 25,
    "total_tokens": 70
  }
}

The content field contains valid JSON that strictly conforms to your schema. Parse it directly in your application without additional validation.

POST/v1/chat/completions (JSON Mode)

Chat Completions with JSON Mode

Generate chat completions that respond with valid JSON (no schema required).

Simple JSON Response

json

{
  "model": "gpt-5.1",
  "messages": [
    {
      "role": "user",
      "content": "Return a JSON object with current weather information for San Francisco"
    }
  ],
  "response_format": {
    "type": "json_object"
  }
}

Use "type": "json_object" for simple JSON responses without a specific schema. The AI will respond with valid JSON but without structural guarantees.

GET/v1/models

List Models

Retrieve a list of all available models.

No parameters required. Simply make the request to this endpoint.

Example Response

json

{
  "models": [
    {
      "id": "gpt-5.1",
      "name": "GPT-5.1",
      "provider": "openai",
      "description": "Most advanced OpenAI model with enhanced reasoning",
      "capabilities": ["text", "code", "reasoning", "vision"],
      "contextLength": "200000",
      "available": true
    }
  ],
  "providers": {
    "openai": true,
    "anthropic": true,
    "google": false,
    "deepseek": true,
    "kimi": false,
    "grok": true
  },
  "defaultModelId": "gpt-5.1"
}

GET/v1/usage

Usage Statistics

Get your API usage statistics and billing information.

Parameter	Type	Required	Description
`period`	`string`	Optional	Optional time window such as "7d", "30d", "90d", "1y" (default: 30d).

Example Response

json

{
  "totalConversations": 42,
  "totalMessages": 310,
  "userMessages": 150,
  "assistantMessages": 160,
  "messagesToday": 12,
  "tokenUsage": 123456,
  "inputTokens": 0,
  "outputTokens": 0,
  "costSavings": 27,
  "modelBreakdown": [
    { "model": "gpt-5.1", "count": 200, "percentage": 65 },
    { "model": "claude-sonnet-4-5", "count": 110, "percentage": 35 }
  ]
}

GET/v1/providers

Supported Providers

Dynamically list all LLM providers currently enabled for your account.

No parameters required. Simply make the request to this endpoint.

Example Response

json

{
  "providers": {
    "openai": true,
    "anthropic": true,
    "google": false,
    "deepseek": true,
    "kimi": false,
    "grok": true
  }
}

GET/v1/providers/{provider}/models

Provider Models

Retrieve only the models for a specific provider (e.g. openai, anthropic, deepseek).

Parameter	Type	Required	Description
`provider`	`string`	Required	Path parameter. One of "openai", "anthropic", "google", "deepseek", "kimi", "grok".

Example Response

json

{
  "provider": "openai",
  "count": 3,
  "models": [
    { "id": "gpt-5.1", "name": "GPT-5.1", "provider": "openai" },
    { "id": "gpt-5", "name": "GPT-5", "provider": "openai" },
    { "id": "gpt-4.1", "name": "GPT-4.1", "provider": "openai" }
  ]
}

GET/v1/models/recommendation

Smart Model Recommendation

Let RyanChat select the best model for you based on cost, speed, or a balanced strategy.

Parameter	Type	Required	Description
`strategy`	`string`	Optional	One of "cheapest", "fastest", or "balanced" (default: balanced).
`provider`	`string`	Optional	Optional provider to constrain the recommendation (e.g. "openai").

Example Response

json

{
  "strategy": "cheapest",
  "provider": "deepseek",
  "model": {
    "id": "deepseek-chat",
    "name": "DeepSeek Chat",
    "provider": "deepseek",
    "tier": "budget",
    "customPricing": {
      "inputCost": 0.15,
      "outputCost": 0.15,
      "currency": "USD"
    },
    "capabilities": ["text", "code", "reasoning"],
    "maxContextLength": 16000
  }
}

Supported Models

Access the most advanced AI models from leading providers through a single API. We continuously add new models and can enable additional ones upon request.

Supported Chat Models

OpenAI Models

GPT-5.1

Latest

Most advanced OpenAI model with enhanced reasoning

Context: 200,000 tokens

Capabilities: Text, Code, Reasoning, Vision

GPT-5

Popular

Next-generation OpenAI model with advanced capabilities

Context: 128,000 tokens

Capabilities: Text, Code, Reasoning, Vision

GPT-4.1

Reliable

Proven OpenAI model with excellent reasoning

Context: 128,000 tokens

Capabilities: Text, Code, Analysis

O1 Models

Reasoning

Specialized models for complex reasoning tasks

Context: Variable

Capabilities: Advanced Reasoning, Math, Science

Supported Image Generation Models

Image models are not currently enabled on this environment. Contact support if you need image generation capabilities enabled for your account.

Need a Specific Model?

Don't see the model you need? We can enable additional models from any supported provider upon request. Contact us to add new models to your account.

Rate Limits & Billing

Rate Limits

10,000 requests per minute
1,000,000 requests per day
10,000,000 tokens per day
Unlimited for Enterprise

Higher limits available upon request for enterprise customers.

Billing & Pricing

Flat 20% margin over underlying LLM provider costsExample: if the upstream model is $1.00 per 1M tokens, RyanChat bills $1.20 per 1M tokens.
Prepaid balance system
Pay only for tokens used
No hidden fees

Volume Discounts: For high-volume or enterprise usage, we can negotiate custom margins and limits. Contact us for enterprise pricing.

Real-time Usage Monitoring

99.9%

API Uptime

< 50ms

Average Latency

24/7

Monitoring

Comprehensive dashboards, real-time alerts, and detailed usage analytics to help you optimize your AI spending and performance.

SDKs & Libraries

Python (REST Example)

Use the RyanChat REST API directly from Python today using requests or httpx. Official SDKs will be added here once released.

Example (requests):

python

import requests

resp = requests.post(
    "https://dev.ryanchat.com/v1/chat/completions",
    headers={"Authorization": "Bearer YOUR_API_KEY"},
    json={
        "model": "gpt-5.1",
        "messages": [{"role": "user", "message": "Hello from Python!"}],
    },
)
print(resp.json())

Node.js (Axios / Fetch)

Call the RyanChat API from Node.js using axios, fetch, or any HTTP client. Official Node.js SDK is planned but not yet released.

Example (axios):

javascript

import axios from "axios";

const resp = await axios.post(
  "https://dev.ryanchat.com/v1/chat/completions",
  {
    model: "gpt-5.1",
    messages: [{ role: "user", message: "Hello from Node!" }],
  },
  {
    headers: { Authorization: "Bearer YOUR_API_KEY" },
  }
);

console.log(resp.data);

REST API

Direct HTTP API access with comprehensive documentation, multiple authentication methods, and works with any programming language.

Features: HTTP/HTTPS, API key auth, rate limiting

bash

curl -X POST https://api.ryanchat.com/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json"

More SDKs Coming Soon

Coming Q1 2025

PHP

Coming Q1 2025

Java

Coming Q2 2025

.NET

Coming Q2 2025

Advanced Multi-Provider Patterns

Leverage RyanChat's multi-LLM infrastructure for advanced use cases that go beyond simple completions. These patterns combine multiple providers and models to achieve superior results.

Hallucination Killer

Run the same prompt through multiple models from different providers (e.g., GPT-5, Claude Opus, Gemini Pro), then compare and validate their responses. Identify inconsistencies and hallucinations by cross-referencing outputs. Perfect for fact-checking, research, and high-stakes decisions.

Use Case: Validate medical advice, legal analysis, or financial recommendations by ensuring multiple top-tier models agree before presenting results to end users.

Consensus GPT

Query 3–5 different models in parallel, aggregate their responses, and synthesize a consensus answer. Use voting, weighted averaging, or a final "referee" model to combine insights. Reduces bias from any single provider and produces more robust, well-rounded responses.

Use Case: Content moderation, sentiment analysis, or generating balanced summaries where multiple perspectives improve accuracy and fairness.

Smart Failover

Automatically switch to a backup model if your primary choice is unavailable, rate-limited, or returns an error. Define fallback chains (e.g., GPT-5 → Claude Opus → Gemini Pro) to ensure 99.9% uptime. RyanChat handles retries and model selection transparently.

Use Case: Mission-critical applications like customer support chatbots or real-time translation services where downtime is not acceptable.

Cost-Aware Routing

Use the /v1/models/recommendation?strategy=cheapest endpoint to automatically route requests to the most cost-effective model that meets your requirements. Balance quality and price by setting constraints (e.g., "cheapest model with 100k+ context").

Use Case: High-volume applications like bulk content generation, data labeling, or automated moderation where cost savings scale significantly.

Hybrid Flows (Multi-Stage Processing)

Chain multiple models together for complex workflows. For example: use a fast, cheap model (DeepSeek) for initial drafts, then refine with a professional model (GPT-5 or Claude Opus) for final polish. Or use a reasoning-focused model (O1) for planning, then a code-specialized model (Codex) for implementation.

Example Flow 1: DeepSeek (draft) → GPT-5 (refinement) → Claude Opus (final review)

Example Flow 2: O1 (reasoning/planning) → Codex (code generation) → GPT-5 (documentation)

Use Case: Content creation pipelines, code generation with review, or research workflows where different models excel at different stages (speed vs. accuracy vs. creativity).

Ready to Build Advanced AI Workflows?

Start experimenting with multi-provider patterns today. No minimum commitment, pay only for what you use.

Get Started