Quickstart
This guide walks you through installing the Fyntune SDK, wrapping your first LLM call, running an eval, and reading the results. Estimated time: 15 minutes.
Prerequisites
- Python 3.9+ or Node.js 18+
- A Fyntune account (sign up free)
- Your Fyntune API key from the dashboard
Step 1: Install the SDK
# Python
pip install fyntune-sdk
# TypeScript / Node.js
npm install @fyntune/sdk
Step 2: Set your API key
Export your API key as an environment variable. Never hard-code it in source files.
export FYNTUNE_API_KEY="ck_your_api_key_here"
Step 3: Wrap your LLM call
Use the @track decorator on any function that calls an LLM. Fyntune captures the inputs and outputs automatically without changing your function's behavior.
from fyntune import track
import openai
@track(feature="summarization", version="v1")
def summarize(text: str) -> str:
response = openai.chat.completions.create(
model="gpt-4o",
messages=[
{"role": "system", "content": "You are a helpful summarization assistant."},
{"role": "user", "content": text}
]
)
return response.choices[0].message.content
Step 4: Make a call
Call your function normally. Fyntune captures the data and runs the eval asynchronously.
result = summarize("The quarterly earnings report showed a 12% increase in revenue...")
print(result)
# Your LLM output appears immediately. Eval runs in background.
Step 5: Read the results
Go to your Fyntune dashboard. You'll see an eval run in progress — it typically completes within 90 seconds. The result shows per-criterion scores for your feature.
Example eval output in the dashboard:
Feature: summarization Version: v1 Run: #0001
Criterion Score Threshold Status
─────────────────────── ──────── ───────── ──────
Semantic Similarity 0.874 0.820 PASS
Factuality 0.921 0.900 PASS
Guardrail Compliance 0.997 0.980 PASS
Response Coherence 0.882 0.850 PASS
Tone Consistency 0.908 0.880 PASS
Overall: PASS ✓ Eval run #0001 complete (67s)
Next steps
- API Reference — full SDK and REST API documentation
- Eval Types — understand all 42 default criteria
- Integrations — add Fyntune to your CI/CD pipeline