Fyntune Documentation
Fyntune's documentation covers everything you need to wire automated LLM eval into your deployment pipeline: SDK setup, CI/CD integration, eval criteria configuration, and the REST API.
The quickstart takes about 15 minutes. The @track decorator wraps your existing LLM calls — no changes to your application logic required.
Where to start
Quickstart
Install the SDK, run your first eval, and read the results. Takes about 15 minutes.
API Reference
Full endpoint documentation with request and response schemas.
Eval Types
The 42 default criteria — what each one measures and when to use it.
Integrations
GitHub Actions, GitLab CI, Vercel, and LLM provider setup guides.
Requirements
Fyntune requires:
- Python 3.9+ or TypeScript/Node.js 18+ for the SDK
- A Fyntune API key (from your dashboard after signup)
- Network access to
api.fyntuneq.comfrom your CI environment
SDK installation
# Python
pip install fyntune-sdk
# TypeScript / Node.js
npm install @fyntune/sdkQuick example
from fyntune import track, FyntuneClient
client = FyntuneClient(api_key="ck_your_api_key")
@track(client=client, feature="summarization", version="v12")
def summarize(text: str) -> str:
# your existing LLM call here
...After wrapping your function, Fyntune captures call data and evaluates it against your configured criteria. Results appear in the dashboard within 90 seconds.