nerif 1.2.0

LLM powered Python

Nerif

Program with both Python and Natural Language.

LLMs can be tricky to work with. They sometimes give overly formal responses or get confused when they're unsure, making them challenging to use in real applications.

That's why we built Nerif - a simple tool that connects LLMs with Python code. We want to make it better than existing tools like Langchain and Dify.

Nerif helps you:

• Control exactly how LLMs work in your code
• Convert LLM responses into usable data formats
• Track performance with metrics like cost and success rate

Our goal is to make LLMs easier to use for developers, turning complex AI capabilities into practical programming tools.

How to install

Pre-requisite

• Python >= 3.9
• Environment variable OPENAI_API_KEY or other LLM API keys, for more details, see here
• Set default model and embedding model with NERIF_DEFAULT_LLM_MODEL and NERIF_DEFAULT_EMBEDDING_MODEL, for more details, see here

Example:

export OPENAI_API_KEY=`your_api_key`
export NERIF_DEFAULT_LLM_MODEL=gpt-4o
export NERIF_DEFAULT_EMBEDDING_MODEL=text-embedding-3-small

Install

pip install nerif

Optional feature groups:

pip install "nerif[img-gen]"
pip install "nerif[asr]"
pip install "nerif[tts]"
pip install "nerif[pydantic]"

QuickStart

from nerif.core import nerif
from nerif.model import SimpleChatModel

model = SimpleChatModel()
# Default model is `gpt-4o`

# Use nerif judge "natural language statement"
if nerif("the sky is blue"):
    print("True")
else:
    # Call a simple model
    print("No", end=", ")
    print(model.chat("what is the color of the sky?"))

v1.1 Features

Streaming

from nerif.model import SimpleChatModel

model = SimpleChatModel()
for chunk in model.stream_chat("Tell me a story"):
    print(chunk, end="", flush=True)

Async Support

import asyncio
from nerif.model import SimpleChatModel

async def main():
    model = SimpleChatModel()
    result = await model.achat("Hello!")
    print(result)

asyncio.run(main())

Pydantic Structured Output

from pydantic import BaseModel
from nerif.model import SimpleChatModel

class City(BaseModel):
    name: str
    country: str
    population: int

model = SimpleChatModel()
city = model.chat("Tell me about Tokyo.", response_model=City)
print(f"{city.name}: {city.population:,}")

Retry Configuration

from nerif.model import SimpleChatModel
from nerif.utils import RetryConfig

model = SimpleChatModel(retry_config=RetryConfig(max_retries=5))

Optional Embedding

from nerif.core import Nerif

# Works without embedding model
judge = Nerif(model="gpt-4o", embed_model=None)
result = judge.judge("the sky is blue")

v1.2.0 Features

Conversation Memory

Sliding window and token-based context management with auto-summarization and persistence.

Enhanced Observability

Latency tracking, cost calculation, success/failure rates, and callback hooks via NerifTokenCounter.

Lightweight RAG

Abstract vector store interface with built-in NumPy implementation for retrieval-augmented generation.

Custom Exceptions

Structured error hierarchy: NerifError, ProviderError, FormatError, ConfigError.

CLI Tools

nerif check        # Verify environment and API keys
nerif test-model   # Test model connectivity
nerif models       # List available models

Type Hints

py.typed marker included for full IDE and mypy support.

Installation

pip install nerif           # Core features
pip install nerif[pydantic] # + Pydantic structured output

Documentation

More detailed documentation is on the official website.

License

Nerif is licensed under the GNU General Public License v3.0.