API Reference

Free LLM API Providers

Groq FREE

Ultra-fast inference on LPU hardware. Free tier includes generous daily token limits.

Llama 3.3 70B Llama 3.1 8B Mixtral 8x7B Gemma 2 9B

Get API Key →

✦

Google Gemini FREE

Google's multimodal AI. Free tier with 15 RPM and 1M tokens/day on Gemini 2.0 Flash.

Gemini 2.0 Flash Gemini 1.5 Pro Gemini 1.5 Flash

Get API Key →

OpenRouter FREE

Access 100+ models through one API. Many models available completely free.

Llama 3.3 70B Mistral 7B Qwen 2.5 72B DeepSeek V3

Get API Key →

Cerebras FREE

Fastest inference in the world on wafer-scale chips. Free tier available.

Llama 3.3 70B Llama 3.1 8B

Get API Key →

Together AI FREE

$1 free credit on signup. Wide model selection with fast inference.

Llama 3.3 70B Qwen 2.5 72B Mixtral 8x7B DeepSeek V3

Get API Key →

🤗

Hugging Face FREE

Free Inference API for thousands of models. Rate-limited but no cost.

Llama 3.3 70B Mistral 7B Phi-3 Zephyr

Get API Key →

☁

Cloudflare AI FREE

10K neurons/day free. Run AI at the edge with zero cold starts.

Llama 3.1 8B Mistral 7B Phi-2 CodeLlama

Get API Key →

SambaNova FREE

Free tier with fast inference on custom RDU chips. No credit card needed.

Llama 3.1 70B Llama 3.1 8B DeepSeek V3

Get API Key →

Mistral AI FREE

European AI lab. Free tier on La Plateforme with Mistral and Codestral models.

Mistral Small Mistral Nemo Codestral Pixtral

Get API Key →

Cohere FREE

Free trial with 100 calls/min. Great for RAG, search, and embeddings.

Command R+ Command R Embed v3

Get API Key →

DeepSeek FREE

Top coding models. Free tier with 500K tokens on signup, then very cheap.

DeepSeek V3 DeepSeek R1 DeepSeek Coder

Get API Key →

Novita AI FREE

Free credits on signup. Wide model selection with competitive pricing.

Llama 3.3 70B DeepSeek V3 Qwen 2.5 72B

Get API Key →

Nebius AI FREE

$1 free credit. Cloud platform with fast GPU inference.

Llama 3.3 70B Mistral 7B Qwen 2.5

Get API Key →

SiliconFlow FREE

Free tier with many open-source models. Popular in Asia.

Qwen 2.5 72B DeepSeek V3 Yi 1.5 34B

Get API Key →

🔥

Fireworks AI FREE

$1 free credit. Fast inference with function calling support.

Llama 3.3 70B Mixtral 8x7B DeepSeek V3

Get API Key →

NVIDIA NIM FREE

1000 free API calls. NVIDIA's optimized inference endpoints.

Llama 3.1 70B Mistral 7B Phi-3

Get API Key →

↑

Upstage FREE

Free tier with Solar models. Strong reasoning and instruction following.

Solar Mini Solar Pro

Get API Key →

CLI Commands

Control Exort from your terminal

💬

Exort chat

Start an interactive chat session with an AI provider.

Exort chat [OPTIONS]

--providerLLM provider (openai, groq, ollama)

--modelModel name

--systemSystem prompt override

--max-tokensMax response tokens (default: 1024)

--temperatureTemperature (default: 0.7)

--streamEnable streaming (default: true)

⚙️

Exort config

Manage Exort configuration settings.

Exort config show # Show current config
Exort config set KEY VALUE # Set a value
Exort config path # Show config file path
Exort config reset # Reset to defaults

🔌

Exort providers

List and test available LLM providers.

Exort providers # List all providers
Exort providers test # Test connectivity

🔧

Exort tools

List and test available agent tools.

Exort tools # List all tools
Exort tools test TOOL_NAME # Test a tool

🚀

Exort serve

Start a local API server (coming soon).

Exort serve --port 8080 --host 0.0.0.0

Python API

Build programmatically with Exort

🤖

Agent

Create an agent and chat with AI, with or without tools.

from Exort import Agent, Config

config = Config()
agent = Agent(config)

# Simple chat
response = await agent.chat("What is 2+2?")

# Chat with tool use
response = await agent.chat("Search the web for AI news")

# Streaming
async for chunk in agent.stream("Tell me a story"):
    print(chunk, end="")

⚙️

Config

Load and manage configuration from files or environment.

from Exort import Config

config = Config()

config.get("provider")  # "groq"
config.set("model", "llama-3.3-70b")

# Load from file
config = Config.from_file("~/.Exort/config.yaml")

# Override with env vars
config = Config.from_env()

🔌

Providers

Use providers directly for custom integrations.

from Exort.providers.groq import GroqProvider

provider = GroqProvider(api_key="gsk_...")

# Non-streaming
response = await provider.chat(
    messages=[{"role": "user", "content": "Hello"}],
    model="llama-3.3-70b-versatile"
)
print(response.content)

# Streaming
async for chunk in provider.stream(messages=[...]):
    print(chunk, end="")

Tools & Memory

Extend Exort with custom tools and persistent memory

🔧

Custom Tools

from Exort.tools.base import tool, ToolRegistry

@tool(
    name="my_tool",
    description="Does something cool",
    parameters={
        "input": {"type": "string"}
    }
)
async def my_tool(input: str) -> str:
    return f"Processed: {input}"

registry = ToolRegistry()
registry.discover()
result = await registry.execute(
    "web_search", query="hello"
)

🧠

Memory Store

Persistent memory for multi-turn conversations.

from Exort.memory.store import MemoryStore

store = MemoryStore()

# Save messages
store.save_message(
    user_id="user1",
    role="user",
    content="Hello"
)

# Get history
history = store.get_history(
    user_id="user1", limit=10
)

# Stats
stats = store.get_stats(user_id="user1")