@voxor/lineage-core

Genetic-algorithm engine for LLM prompt optimization. Evaluates candidate prompts against a test set across multiple LLM providers (OpenAI, Anthropic, Gemini, OpenRouter, Groq), evolves them with mutation/crossover/meta-prompting operators, and scores fitness on quality, safety, cost, latency, and stability.

This is the embeddable engine. Most users want:

• @voxor/lineage-cli — command-line runner (npx lineage)
• Lineage desktop app — visual evolution graph

Programmatic use

The host injects platform services before starting an evaluation:

import {
  setStore,
  setSendUpdate,
  initializeDatabase,
  startEvaluation,
} from '@voxor/lineage-core';

See the repository for full documentation.