Skip to main content

​
Overview

Access 570+ AI models from leading providers through a unified interface. Splox supports major LLM providers for text generation, code completion, reasoning, and multimodal tasks including image, video, and audio generation.

570+ AI Models

400+ LLMs + 170+ generative AI models

6 Major Providers

OpenAI, Anthropic, Google, Perplexity, OpenRouter

Multimodal AI

Text, image, video, audio, and speech

Cost Optimization

Choose models based on performance and cost

​
API Key Options

Splox offers two ways to access AI providers:
Let Splox handle API keys and infrastructure for you.Benefits:
  • No need to sign up for provider accounts
  • No API key management required
  • Instant access to all providers
  • Single consolidated billing through Splox
How it works:
  • When creating an integration, set use_internal_service: true
  • Splox uses its own API credentials
  • You are charged for token usage + commission in a single bill
Use your own API keys from providers for direct billing control.Benefits:
  • Direct relationship with AI provider
  • More control over API key management
  • Track usage directly in provider dashboard
How it works:
  • When creating an integration, set use_internal_service: false
  • Provide your own API key from the provider
  • Provider bills you directly for token usage
  • Splox charges a small commission for platform features
Commission Model: Splox charges a commission on all AI provider usage, whether you use Splox Services or BYOK. This enables platform features like usage tracking, workflow orchestration, and support.

​
OpenAI

The industry leader in large language models and generative AI.
  • GPT-5 - Next generation flagship (Pro, Mini, Nano, Codex variants)
  • GPT-4.1 - Latest GPT-4 series (Standard and Mini variants)
  • o1, o3, o4 - Advanced reasoning models
  • GPT-4o - Optimized for speed and cost
  • GPT-4o-mini - Fast and affordable
  • Codex - Code generation specialist

​
Available Actions

Generate text responses using GPT models.Parameters:
  • model - GPT-5, GPT-4.1, GPT-4o, GPT-4o-mini, o1, o3, o4, etc.
  • messages - Array of conversation messages
  • temperature - Randomness (0-2, default: 0.7)
  • max_tokens - Maximum response length
  • top_p - Nucleus sampling (0-1)
  • frequency_penalty - Reduce repetition (-2 to 2)
  • presence_penalty - Encourage new topics (-2 to 2)
  • stream - Enable streaming responses
Returns:
  • content - Generated text
  • finish_reason - Completion reason
  • usage - Token usage statistics
Convert text to natural-sounding speech.Models:
  • gpt-4o-mini TTS
Parameters:
  • text - Text to convert
  • voice - Voice selection
  • speed - Speech speed (0.25-4.0)
Returns:
  • audio_url - URL to generated audio file
Transcribe audio to text using Whisper.Models:
  • Whisper
  • GPT-4o-mini Transcribe
Parameters:
  • audio_url - URL to audio file
  • language - Audio language (optional)
  • prompt - Context to guide transcription
Returns:
  • text - Transcribed text
  • language - Detected language
Generate vector embeddings for semantic search.Models:
  • text-embedding-3-large
  • text-embedding-3-small
Parameters:
  • input - Text to embed
  • model - Embedding model
Returns:
  • embeddings - Vector array
  • dimensions - Vector dimensions
Generate images from text descriptions.Models:
  • GPT Text To Image 1
Parameters:
  • prompt - Image description
  • size - 1024x1024, 1792x1024, 1024x1792
  • quality - standard, hd
  • style - natural, vivid
  • num_images - Number to generate
Returns:
  • images - Array of image URLs
  • revised_prompt - Enhanced prompt used
Edit existing images with AI.Models:
  • GPT Image To Image 1
Parameters:
  • image_url - Source image URL
  • prompt - Edit description
  • mask_url - Edit area mask (optional)
  • size - Output size
  • num_images - Number of variations
Returns:
  • images - Array of edited image URLs
Generate videos from text prompts.Models:
  • Sora 2
  • Sora 2 Pro
  • Sora 2 Remix
  • Sora 2 Pro Remix
Parameters:
  • prompt - Video description
  • duration - Video length in seconds
  • resolution - 1080p, 720p, 480p
  • aspect_ratio - 16:9, 9:16, 1:1
Returns:
  • video_url - Generated video URL
  • thumbnail_url - Video thumbnail

​
Setup

  1. Get API key from OpenAI Platform
  2. Add integration in Splox: Settings > Integrations > OpenAI
  3. Enter your API key

​
Anthropic

Advanced AI models known for safety and superior reasoning capabilities.
  • Claude 4.5 Sonnet - Latest flagship
  • Claude 4.5 Haiku - Fast and affordable
  • Claude 4.1 Opus - Most capable
  • Claude 4.0 Opus - Highly capable
  • Claude 3.7 Sonnet - Balanced performance

​
Key Features

  • Extended context windows (up to 200K tokens)
  • Superior reasoning capabilities
  • Built-in safety features
  • Computer use capabilities

​
Available Actions

Generate text responses using Claude models.Parameters:
  • model - Claude 4.5 Sonnet, Claude 4.0 Opus, Claude 3.7 Sonnet, etc.
  • messages - Array of conversation messages
  • temperature - Randomness (0-1, default: 0.7)
  • max_tokens - Maximum response length
  • top_p - Nucleus sampling (0-1)
  • stream - Enable streaming responses
Returns:
  • content - Generated text
  • finish_reason - Completion reason
  • usage - Token usage statistics

​
Setup

  1. Get API key from Anthropic Console
  2. Add integration in Splox: Settings > Integrations > Anthropic
  3. Enter your API key

​
Google

Google’s cutting-edge AI models with multimodal capabilities.
  • Gemini 2.5 Pro - Most capable
  • Gemini 2.5 Flash - Fast and efficient
  • Gemini 2.5 Flash Lite - Ultra-fast
  • Gemma 3 - Open source (27B, 12B, 4B variants)

​
Strengths

  • Multimodal understanding (text, images, video)
  • Long context windows
  • Fast inference
  • Thinking/reasoning modes

​
Available Actions

Generate text responses using Gemini models.Parameters:
  • model - Gemini 2.5 Pro, Gemini 2.5 Flash, Gemini 2.5 Flash Lite, etc.
  • messages - Array of conversation messages
  • temperature - Randomness (0-2, default: 0.7)
  • max_tokens - Maximum response length
  • top_p - Nucleus sampling (0-1)
  • stream - Enable streaming responses
Returns:
  • content - Generated text
  • finish_reason - Completion reason
  • usage - Token usage statistics
Convert text to natural-sounding speech.Models:
  • Gemini 2.5 Pro TTS
  • Gemini 2.5 Flash TTS
Parameters:
  • text - Text to convert
  • voice - Voice selection
  • speed - Speech speed (0.25-4.0)
Returns:
  • audio_url - URL to generated audio file
Generate vector embeddings for semantic search.Models:
  • text-embedding-004
Parameters:
  • input - Text to embed
  • model - Embedding model
Returns:
  • embeddings - Vector array
  • dimensions - Vector dimensions

​
Setup

  1. Get API key from Google AI Studio
  2. Add integration in Splox: Settings > Integrations > Google
  3. Enter your API key

​
Additional Providers

Open-source models with excellent performance. Available via OpenRouter.Popular Models:
  • Llama 4 Maverick, Llama 4 Scout
  • Llama 3.3 70B, Llama 3.2 (vision)
  • Llama Guard (safety)
Strengths:
  • Open source and customizable
  • Cost-effective
  • Strong coding capabilities
  • Multiple size variants
Specialized in reasoning and code generation. Available via OpenRouter.Popular Models:
  • DeepSeek R1 - Advanced reasoning
  • DeepSeek V3 - General purpose
  • DeepSeek Prover V2 - Mathematical proofs
  • DeepSeek Coder - Code specialist
Use Cases:
  • Complex reasoning tasks
  • Mathematical proofs
  • Code generation and analysis
  • Chain-of-thought reasoning
European AI provider with competitive models. Available via OpenRouter.Popular Models:
  • Mistral Large - Flagship
  • Mistral Medium - Balanced
  • Mistral Small - Fast
  • Codestral - Code specialist
  • Pixtral - Multimodal
Powerful multilingual models. Available via OpenRouter.Popular Models:
  • Qwen3 235B - Largest
  • Qwen3 Max - Best performance
  • QwQ 32B - Reasoning
  • Qwen2.5 VL - Vision-language
Strengths:
  • Excellent multilingual support
  • Strong reasoning capabilities
  • Vision understanding
  • Cost-effective
Real-time conversational AI. Available via OpenRouter.Popular Models:
  • Grok 4, Grok 3 Beta
Features:
  • Real-time information
  • Conversational AI
Enterprise-focused AI models. Available via OpenRouter.Popular Models:
  • Command R+ (RAG specialist)
Features:
  • Multilingual support
  • Enterprise features
  • RAG optimization
AWS-integrated AI models. Available via OpenRouter.Popular Models:
  • Nova models
Features:
  • AWS integration
  • Enterprise support
GPU-accelerated inference. Available via OpenRouter.Popular Models:
  • Llama Nemotron
Features:
  • Optimized inference
  • GPU acceleration
Unified gateway to 300+ AI models from multiple providers.Features:
  • Access to 300+ models from all major providers
  • Automatic routing and fallback
  • Cost optimization
  • Single API for all models
Providers Accessible:
  • Meta, DeepSeek, Mistral, Qwen, xAI
  • Cohere, Amazon Bedrock, NVIDIA
  • Microsoft, Google, Anthropic
  • And many more
Setup:
  1. Get API key from OpenRouter
  2. Add integration in Splox
  3. Access any model with openrouter/ prefix

​
Model Selection Guide

Best for everyday conversational AI tasks.Recommended Models:
  • GPT-5 (best quality)
  • Claude 4.5 Sonnet (safety-focused)
  • Gemini 2.5 Flash (fast & affordable)
  • Llama 3.3 70B (open source)
For complex problem-solving and analysis.Recommended Models:
  • Claude 4.1 Opus (best reasoning)
  • o4, o3 (OpenAI)
  • Claude 4.0 Opus
  • DeepSeek R1
  • QwQ 32B
For programming and technical tasks.Recommended Models:
  • Claude 4.5 Sonnet (best for coding)
  • Claude 4.1 Opus (best for coding)
  • Claude 4.0 Opus (best for coding)
  • GPT-5 Codex
  • DeepSeek Coder
  • Codestral
For image understanding and analysis.Recommended Models:
  • GPT-4o (vision)
  • Gemini 2.5 Pro
  • Qwen VL
For high-volume, cost-sensitive applications.Recommended Models:
  • Claude 4.5 Haiku (fast & cheap)
  • GPT-4o-mini
  • Gemini 2.5 Flash Lite
  • Groq (Llama)
For creating images from text.Recommended Models:
  • GPT Text To Image 1 (OpenAI)
For creating videos from text/images.Recommended Models:
  • Sora 2 Pro (OpenAI)
For TTS, voice cloning, music.Recommended Models:
  • Whisper (OpenAI)

​
Troubleshooting

Solution:
  • Add delays between requests
  • Implement exponential backoff
  • Use multiple API keys
  • Upgrade to higher tier
Solution:
  • Summarize or truncate input
  • Use models with larger context
  • Split into multiple requests
  • Implement sliding window
Solution:
  • Improve prompt clarity
  • Add examples
  • Adjust temperature
  • Try different model
Solution:
  • Use cheaper models
  • Reduce token usage
  • Implement caching
  • Optimize prompts

​
Next Steps

LLM Node

Learn about the LLM node

Tool Node

Combine LLMs with tools

Memory Node

Add conversation memory

Prompt Engineering

Master prompt writing