Promptify

Task-based NLP engine with Pydantic structured outputs, built-in evaluation, and LiteLLM as the universal LLM backend. Think "scikit-learn for LLM-powered NLP".

Installation

With pip

Requires Python 3.9+.

pip install promptify

or

pip install git+https://github.com/promptslab/Promptify.git

For evaluation metrics support:

pip install promptify[eval]

Quick Tour

3-Line NER

from promptify import NER
ner = NER(model="gpt-4o-mini", domain="medical") result = ner("The patient is a 93-year-old female with a medical history of chronic right hip pain, osteoporosis, hypertension, depression, and chronic atrial fibrillation admitted for evaluation and management of severe nausea and vomiting and urinary tract infection")

Output:

NERResult(entities=[     Entity(text="93-year-old", label="AGE"),     Entity(text="chronic right hip pain", label="CONDITION"),     Entity(text="osteoporosis", label="CONDITION"),     Entity(text="hypertension", label="CONDITION"),     Entity(text="depression", label="CONDITION"),     Entity(text="chronic atrial fibrillation", label="CONDITION"),     Entity(text="severe nausea and vomiting", label="SYMPTOM"),     Entity(text="urinary tract infection", label="CONDITION"), ])

Classification

from promptify import Classify
clf = Classify(model="gpt-4o-mini", labels=["positive", "negative", "neutral"]) result = clf("Amazing product! Best purchase I've ever made.") 
Classification(label="positive", confidence=0.95)

Question Answering

from promptify import QA
qa = QA(model="gpt-4o-mini") answer = qa("Einstein was born in Ulm in 1879.", question="Where was Einstein born?") 
Answer(answer="Ulm", evidence="Einstein was born in Ulm", confidence=0.98)

Custom Task with Any Pydantic Schema

from promptify import Task
from pydantic import BaseModel
class MovieReview(BaseModel):     sentiment: str     rating: float     key_themes: list[str]
task = Task(model="gpt-4o", output_schema=MovieReview, instruction="Analyze this movie review.") review = task("Nolan's best work. Stunning visuals but the plot drags.") 
MovieReview(sentiment="mostly positive", rating=7.5, key_themes=["visuals", "pacing"])

Any Provider - Just Change the Model String

ner_openai = NER(model="gpt-4o-mini")
ner_claude = NER(model="claude-sonnet-4-20250514")
ner_local  = NER(model="ollama/llama3")

Batch Processing

results = ner.batch(["text1", "text2", "text3"], max_concurrent=10)

Async Support

result = await ner.acall("Patient has diabetes")

Built-in Evaluation

from promptify.eval import evaluate
scores = evaluate(task=ner, dataset=labeled_data, metrics=["precision", "recall", "f1"]) 
{"precision": 0.92, "recall": 0.88, "f1": 0.90}

Features

• 2-3 lines of code for any NLP task -no training data required
• Pydantic structured outputs -type-safe results, not raw strings
• Any LLM provider via LiteLLM -OpenAI, Anthropic, Google, Ollama, Azure, and 100+ more
• Built-in tasks -NER, Classification (binary/multiclass/multilabel), QA, Summarization, Relation Extraction, SQL Generation, and more
• Custom tasks -bring your own Pydantic schema for any structured output
• Few-shot examples -easily add examples to improve accuracy
• Domain specialization -pass domain="medical" or any domain for context-aware prompts
• Batch processing -async concurrency under the hood for processing multiple texts
• Async support -native await support with acall()
• Evaluation framework -precision, recall, F1, accuracy, exact match, ROUGE metrics
• Safe parser -fallback JSON completion for providers without native structured outputs (no eval())
• Cost tracking -built-in token usage and cost monitoring via getcostsummary()

Supported NLP Tasks

| Task | Class | Output Schema | |------|-------|---------------| | Named Entity Recognition | NER | NERResult (list of Entity) | | Binary Classification | Classify | Classification | | Multiclass Classification | Classify | Classification | | Multilabel Classification | Classify(multi_label=True) | MultiLabelResult | | Question Answering | QA | Answer | | Summarization | Summarize | Summary | | Relation Extraction | ExtractRelations | ExtractionResult | | Tabular Extraction | ExtractTable | ExtractionResult | | Question Generation | GenerateQuestions | list of GeneratedQuestion | | SQL Generation | GenerateSQL | SQLQuery | | Text Normalization | NormalizeText | normalized text | | Topic Modelling | ExtractTopics | list of topics | | Custom Task | Task | any Pydantic BaseModel |

Community

@misc{Promptify2022,
  title = {Promptify: Structured Output from LLMs},
  author = {Pal, Ankit},
  year = {2022},
  howpublished = {\url{https://github.com/promptslab/Promptify}},
  note = {Prompt-Engineering components for NLP tasks in Python}
}

💁 Contributing

We welcome any contributions to our open source project, including new features, improvements to infrastructure, and more comprehensive documentation. Please see the contributing guidelines