What does the Lifeboat prompt chain do?

Lifeboat is a scaffolded promptfoo example that provides a starting point for evaluating AI models using both OpenAI and Replicate API tokens configured in the same promptfooconfig.yaml file, with a standard init/eval/view workflow.

Do I need to add my own test cases to use Lifeboat?

Yes, Lifeboat does not ship with pre-built test cases—you must define your own test cases and assertions in the promptfooconfig.yaml file.

Who should use Lifeboat?

Teams that want a minimal promptfoo starting point for building their own evaluation spanning both OpenAI and Replicate providers rather than starting from a blank configuration.

Prompt Chain

Summarize Web Content with Replicate

Name: Lifeboat
Availability: OnlineOnly
Author: Promptfoo

Minimal promptfoo scaffold for evaluating models across OpenAI and Replicate providers together.

Copy chain

Works with replicate

Promptfoo

Own this? Claim it

Spark score

out of 100

Updated last month

Version code-scan-action-0.1

Add to Favorites

Why it matters

Leverage the Replicate API to process and summarize web content, extracting key information and providing concise summaries.

Outcomes

What it gets done

Fetch and process content from specified web sources.

Extract relevant data points and key themes.

Generate concise summaries of the extracted information.

Integrate with Replicate for AI model execution.

Install

Add it to your toolbox

Run in your project directory:

curl -fsSL https://spark.entire.vc/get/pfoo-lifeboat | bash

Steps

Steps in the chain

Initialize example project

Navigate to project directory

Set API tokens

Add test cases

Run evaluation

View results

Overview

Lifeboat

A minimal promptfoo scaffold intended for evaluating models across both OpenAI and Replicate providers, with test cases to be defined in promptfooconfig.yaml. Use as a starting point when setting up an eval that spans both OpenAI and Replicate providers rather than a blank promptfoo config.

What it does

This is a scaffolded promptfoo example ('Replicate Lifeboat') for evaluating models using both OpenAI and Replicate API tokens configured in the same promptfooconfig.yaml. It provides the standard init/eval/view flow as a starting point.

When to use - and when NOT to

Use this as a bare starting scaffold when you need to set up an eval spanning both OpenAI and Replicate providers and will define the actual test cases and assertions yourself in promptfooconfig.yaml. It does not ship pre-built test cases - those must be added by the user.

Inputs and outputs

Set OpenAI and Replicate API tokens in promptfooconfig.yaml, add test cases, then scaffold and run:

npx promptfoo@latest init --example provider-replicate/lifeboat
cd provider-replicate/lifeboat

followed by promptfoo eval to run the evaluation and promptfoo view to inspect the results.

Integrations

Scaffolds a promptfoo config intended to evaluate models across both OpenAI and Replicate providers within one config file.

Who it's for

Teams that want a minimal promptfoo starting point for building their own eval spanning both OpenAI and Replicate providers rather than starting from a blank config.

Source README

provider-replicate/lifeboat (Replicate Lifeboat)

You can run this example with:

npx promptfoo@latest init --example provider-replicate/lifeboat
cd provider-replicate/lifeboat

Usage

To get started, set your OpenAI and Replicate API tokens in promptfooconfig.yaml.

Then, add your test cases.

Then run:

promptfoo eval

Afterwards, you can view the results by running promptfoo view

FAQ

Common questions

Discussion