New
Overview Tzafon provides an OpenAI-compatible chat completions API. You can use the standard OpenAI SDK with Tzafon’s models by simply changing the base URL.
Use your Tzafon API key (TZAFON_API_KEY) with the OpenAI SDK—no separate authentication required.
Quick Start from openai import OpenAI client = OpenAI( api_key = "sk_your_tzafon_api_key" , base_url = "https://api.tzafon.ai/v1" ) response = client.chat.completions.create( model = "tzafon.sm-1" , messages = [ { "role" : "system" , "content" : "You are a helpful assistant." }, { "role" : "user" , "content" : "Hello!" } ] ) print (response.choices[ 0 ].message.content)
Available Models Model Description Best For tzafon.sm-1Small, fast model General tasks, quick responses tzafon.northstar-cua-fastComputer-use optimized Browser/desktop automation agents
Use GET /v1/models to retrieve the current list of available models.
Endpoints Chat Completions POST https://api.tzafon.ai/v1/chat/completions
Create a chat completion with conversation history.
Request Body: Parameter Type Required Description modelstring Yes Model ID (e.g., tzafon.sm-1) messagesarray Yes Conversation messages temperaturenumber No Sampling temperature (0-2). Default: 1 max_tokensnumber No Maximum tokens to generate streamboolean No Enable streaming responses stoparray No Stop sequences
Message Format: { "role" : "system" | "user" | "assistant" , "content" : "Message text" }
Streaming Enable real-time token streaming for responsive UIs:
from openai import OpenAI client = OpenAI( api_key = "sk_your_tzafon_api_key" , base_url = "https://api.tzafon.ai/v1" ) stream = client.chat.completions.create( model = "tzafon.sm-1" , messages = [{ "role" : "user" , "content" : "Write a short poem" }], stream = True ) for chunk in stream: if chunk.choices[ 0 ].delta.content: print (chunk.choices[ 0 ].delta.content, end = "" )
List Models GET https://api.tzafon.ai/v1/models
Returns the list of available models.
from openai import OpenAI client = OpenAI( api_key = "sk_your_tzafon_api_key" , base_url = "https://api.tzafon.ai/v1" ) models = client.models.list() for model in models.data: print (model.id)
Completions (Legacy) POST https://api.tzafon.ai/v1/completions
For text completion without chat format. Parameters are similar to chat completions, but uses a prompt string instead of messages.
Embeddings POST https://api.tzafon.ai/v1/embeddings
Generate vector embeddings for text.
Configuration Options Temperature Presets Preset Temperature Max Tokens Use Case Creative 1.0 2048 Creative writing, brainstorming Balanced 0.7 1024 General tasks Precise 0.3 512 Factual responses, code
Example with Options response = client.chat.completions.create( model = "tzafon.sm-1" , messages = [{ "role" : "user" , "content" : "Explain quantum computing" }], temperature = 0.7 , max_tokens = 1024 , stop = [ "<|im_end|>" , "<|end_of_text|>" ] )
Error Handling The API returns standard HTTP status codes:
Status Description 200 Success 400 Bad request (invalid parameters) 401 Unauthorized (invalid API key) 429 Rate limit exceeded 500+ Server error
from openai import OpenAI, APIError, RateLimitError, AuthenticationError client = OpenAI( api_key = "sk_your_tzafon_api_key" , base_url = "https://api.tzafon.ai/v1" ) try : response = client.chat.completions.create( model = "tzafon.sm-1" , messages = [{ "role" : "user" , "content" : "Hello" }] ) except AuthenticationError: print ( "Invalid API key" ) except RateLimitError: print ( "Rate limit exceeded - retry later" ) except APIError as e: print ( f "API error: { e } " )
Pricing LLM API usage is billed per million tokens:
Model Input Tokens Output Tokens tzafon.sm-1$0.20 / 1M tokens $0.30 / 1M tokens tzafon.northstar-cua-fast$0.30 / 1M tokens $1.00 / 1M tokens
Using with Computer Automation Combine the LLM API with browser automation for AI-powered workflows:
from openai import OpenAI from tzafon import Computer # LLM client llm = OpenAI( api_key = "sk_your_tzafon_api_key" , base_url = "https://api.tzafon.ai/v1" ) # Computer client computer_client = Computer() with computer_client.create( kind = "browser" ) as computer: computer.navigate( "https://example.com" ) screenshot = computer.screenshot() # Use LLM to analyze the page response = llm.chat.completions.create( model = "tzafon.northstar-cua-fast" , messages = [ { "role" : "system" , "content" : "You are a web automation assistant." }, { "role" : "user" , "content" : f "I'm on a webpage. What should I click next? Screenshot: { screenshot } " } ] ) print (response.choices[ 0 ].message.content)
Next Steps
Quickstart Get started with Tzafon
Computer Instances Browser and desktop automation
