ucbepicucbepicOfficialยท1 Agent Skills Included

docetl

Build LLM pipelines that extract and summarize unstructured data

Builds and runs LLM-powered pipelines that extract, classify, filter, and summarize large collections of documents and text data. Eliminates hand-written LLM calls, manual prompt tuning, and messy one-off scripts by orchestrating map, reduce, filter, and resolve operations automatically. Optimizes pipelines for cost and accuracy by rewriting prompts, swapping models, and replacing subtasks with code, returning clean queryable tables.
npx skills add ucbepic/docetl --all -g -y
Available:

Guides the AI agent through an interactive, step-by-step process for designing, testing, and implementing new pipeline rewrite directives in the reasoning optimizer.

All Skills in This Repository (1)

Pure Emerald Level Indicators

Frequently Asked Questions

FAQPage Schema
How to install DocETL?โ–ผ

Run `npx skills add ucbepic/docetl --all -g -y` in your terminal to install all tools in this suite globally.

What is DocETL used for?โ–ผ

DocETL processes large collections of documents with LLMs using simple natural-language operations like map, filter, and reduce. It turns unstructured text into clean, structured tables you can query.

Do I need to write code to use DocETL?โ–ผ

No. You can declare pipelines in a simple YAML file or describe your task in plain English and let your AI agent build the pipeline for you.

Does DocETL work with Claude Code?โ–ผ

Yes. DocETL ships a SKILL.md that works with Claude Code and other compatible agents, so the agent can build, run, and optimize pipelines from your plain-English requests.

How does DocETL reduce LLM costs?โ–ผ

Its optimizer automatically rewrites prompts, picks cheaper models for simple steps, and replaces LLM subtasks with deterministic code where possible.

Related Repositories in Data & Analytics

View All in Data & Analyticsโ†’