tenro 0.1.5

A modern multi-agent AI testing framework

Tenro

A modern, provider-agnostic simulation engine to safely test AI agents.

Verify multi-agent workflows and tool usage without burning tokens.

Simulate everything. Trust your agents.

Install

pip install tenro
# or
uv add tenro

Quick Start

# myapp/agent.py
from tenro import link_tool, link_llm

@link_tool("search")
def search(query: str) -> str:
    return external_api.search(query)

@link_llm("openai")
def call_llm(prompt: str) -> str:
    return openai.chat.completions.create(...)

# tests/test_agent.py
def test_agent(construct):
    construct.simulate_tool("search", result=["Simulated Doc"])    
    construct.simulate_llm(provider="openai", response="Done")
    
    agent.run("Hello")
    
    construct.verify_tool("search", query="Secret Docs", times=1)
    construct.verify_llm(times=1)

No simulations to configure, no expensive API calls, no flaky tests.

Why Tenro?

• Zero API calls — Tests run instantly, no rate limits or costs
• Full control — Simulate LLM responses, test edge cases reliably
• Ship with confidence — Verify agent behaviour, not just their final response
• pytest-native — Drop-in fixture or standalone, works with your existing setup
• Provider-aware — Simulates OpenAI, Anthropic, Gemini with real response shapes

Before / After

Without Tenro — manual simulations, helper functions, boilerplate

# test_helpers.py - you write and maintain this
def simulate_llm_response(content=None, tool_call=None):
    if tool_call:
        message = ChatCompletionMessage(
            role="assistant", content=None,
            tool_calls=[ChatCompletionMessageToolCall(
                id="call_abc", type="function",
                function=Function(name=tool_call["name"], arguments=json.dumps(tool_call["args"]))
            )]
        )
    else:
        message = ChatCompletionMessage(role="assistant", content=content, tool_calls=None)
    return ChatCompletion(
        id="chatcmpl-123", created=0, model="gpt-5", object="chat.completion",
        choices=[Choice(index=0, finish_reason="stop", message=message)]
    )

# test_agent.py
@patch("myapp.tools.get_weather")
@patch("openai.chat.completions.create")
def test_agent(simulated_llm, simulated_weather):
    simulated_weather.return_value = {"temp": 72, "condition": "sunny"}
    simulated_llm.side_effect = [
        simulate_llm_response(tool_call={"name": "get_weather", "args": {"city": "Paris"}}),
        simulate_llm_response(content="It's 72°F and sunny in Paris."),
    ]
    result = my_agent.run("Weather in Paris?")
    assert result == "It's 72°F and sunny in Paris."
    simulated_weather.assert_called_once_with(city="Paris")

With Tenro:

def test_agent(construct):
    construct.simulate_tool("get_weather", result={"temp": 72, "condition": "sunny"})
    construct.simulate_llm(provider="openai", responses=[
        {"tools": [{"name": "get_weather", "arguments": {"city": "Paris"}}]},
        "It's 72°F and sunny in Paris.",
    ])

    my_agent.run("Weather in Paris?")

    construct.verify_agent("WeatherAgent", output_contains="72°F and sunny")
    construct.verify_tool("get_weather", city="Paris")

No simulations. No helpers. Just behavior.

How It Works

Tenro's Construct is a simulation environment for your AI agents. Link your functions with decorators, then test with full control:

from tenro import link_agent, link_llm, link_tool

@link_agent("Manager")
def manager(task: str) -> str:
    docs = search(task)
    return summarize(docs)

@link_tool("search")
def search(query: str) -> list[str]:
    return external_search_api(query)

@link_llm("openai", model="gpt-5")
def summarize(docs: list[str]) -> str:
    return openai.chat.completions.create(...)

In tests, the construct fixture intercepts these calls and applies your simulations.

Trace Output

Enable trace visualization to debug agent execution:

Set TENRO_TRACE=true in your .env or run TENRO_TRACE=true pytest

🤖 SupportAgent
   ├─ → user: "My order #12345 hasn't arrived"
   │
   ├─ 🧠 claude-sonnet-4-5
   │     ├─ → prompt: "Help customer: My order #12345 hasn't arrived"
   │     └─ ← tool_call: lookup_order(order_id='12345')
   │
   ├─ 🔧 lookup_order
   │     ├─ → order_id='12345'
   │     └─ ← {'status': 'shipped', 'eta': '2025-01-02'}
   │
   ├─ 🧠 claude-sonnet-4-5
   │     ├─ → prompt: "Tool result: {'status': 'shipped', ...}"
   │     └─ ← "Your order has shipped and will arrive by Jan 2nd!"
   │
   └─ ← "Your order has shipped and will arrive by Jan 2nd!"

────────────────────────────────────────────────────────────────
Summary: 1 agent | 2 LLM calls | 1 tool call | Total: 1.24s

LLM Provider Support

Provider	Status
OpenAI	✅
Anthropic	✅
Gemini	✅
Others	🚧 Coming soon

Compatibility

• Python 3.11+
• pytest 7.0+

Contributing

Thanks for your interest in contributing!

We are currently in the early stages of development and are focused on stabilizing the core API. Because of this, we aren't accepting external Pull Requests just yet.

However, your support is incredibly valuable to us. You can help us right now by:

• Starring the Repository ⭐️: This helps others discover the project and lets you track when we open up for code contributions.
• Reporting Bugs: If something breaks, let us know.
• Suggesting Features: Have an idea on how to make this better? Tell us!
• Asking Questions: We are happy to discuss the roadmap and usage.

Please use GitHub Issues for discussions and reports.

License

Apache 2.0

Support

• Issues: GitHub Issues
• Email: [e-mail]