Heabsy

Use Heabsy's OpenAI-compatible API for open models with PraisonAI Agents

Models

Heabsy is an OpenAI-compatible inference API for open models. Heabsy states that models in its EEA tier run on dedicated GPUs in the EEA and that prompts and outputs are not stored (see its zero data retention page and DPA). Models routed through third parties are labelled as such in the catalog.

Pass a bare model id with base_url="https://api.heabsy.com/v1" — PraisonAI routes it as openai/<model>. See OpenAI-Compatible Endpoints.

Recommended Models:

  • Qwen3.8 27B: qwen38 (524,288-token context, up to 32,768 output tokens, vision input, tool calling, structured output)

Python

python
# export HEABSY_API_KEY=your-api-key
import os
from praisonaiagents import Agent

agent = Agent(
    instructions="You are a helpful assistant",
    llm="qwen38",
    base_url="https://api.heabsy.com/v1",
    api_key=os.environ["HEABSY_API_KEY"]
)
agent.start("Explain neural networks")

With Tools

python
# export HEABSY_API_KEY=your-api-key
import os
from praisonaiagents import Agent

def search_web(query: str) -> str:
    """Search the web for information."""
    return f"Search results for: {query}"

agent = Agent(
    instructions="You are a research assistant",
    llm="qwen38",
    base_url="https://api.heabsy.com/v1",
    api_key=os.environ["HEABSY_API_KEY"],
    tools=[search_web]
)
agent.start("What's happening in AI today?")

Multi-Agent

python
# export HEABSY_API_KEY=your-api-key
import os
from praisonaiagents import Agent, Task, AgentTeam

researcher = Agent(
    instructions="You research topics thoroughly",
    llm="qwen38",
    base_url="https://api.heabsy.com/v1",
    api_key=os.environ["HEABSY_API_KEY"]
)
writer = Agent(
    instructions="You write clear summaries",
    llm="qwen38",
    base_url="https://api.heabsy.com/v1",
    api_key=os.environ["HEABSY_API_KEY"]
)

task1 = Task(description="Research quantum computing", agent=researcher)
task2 = Task(description="Summarize findings", agent=writer)

agents = AgentTeam(agents=[researcher, writer], tasks=[task1, task2])
agents.start()

CLI

bash
export HEABSY_API_KEY=your-api-key

# Basic prompt
OPENAI_API_KEY="$HEABSY_API_KEY" OPENAI_API_BASE=https://api.heabsy.com/v1 \
  python -m praisonai "Explain AI" --llm openai/qwen38

# Run agents.yaml
OPENAI_API_KEY="$HEABSY_API_KEY" OPENAI_API_BASE=https://api.heabsy.com/v1 \
  python -m praisonai

The OpenAI variables are set for each command only. Exporting them would send every OpenAI-based tool in that shell to Heabsy.

YAML

yaml
# Run with OPENAI_API_KEY and OPENAI_API_BASE set as in the CLI section
framework: praisonai
topic: Research AI developments
agents:
  researcher:
    role: AI Researcher
    goal: Find latest AI developments
    instructions: You are an expert AI researcher
    llm:
      model: openai/qwen38
    tasks:
      research_task:
        description: Research the latest developments in AI
        expected_output: Comprehensive AI research report

  writer:
    role: Technical Writer
    goal: Create clear documentation
    instructions: You write clear technical content
    llm:
      model: openai/qwen38
    tasks:
      write_task:
        description: Write a summary of the research
        expected_output: Well-written technical summary

Resources