# Introducing the Parallel CLI

## Why a CLI?

The most capable autonomous agents today operate in the terminal, from OpenClaw and Claude Code to custom pipelines built on tool-use frameworks. They reason, write code, execute commands, and iterate. When these agents need web data, they shouldn't have to drop into an SDK or construct HTTP requests. They should run a command.

Parallel CLI is built to be intuitive to agents. Every command accepts _--json _for structured output that agents can parse directly. Every long-running operation supports _--no-wait_ for async execution with separate status polling. And every command works non-interactively, so agents can compose Parallel’s products together into multi-step pipelines without human intervention.

We developed the CLI alongside Parallel Agent Skills[Parallel Agent Skills](https://github.com/parallel-web/parallel-agent-skills), a set of structured skill definitions that teach agents when and how to use each command. Together, the CLI and Agent Skills give your agent a complete, self-contained web intelligence toolkit for finding anything on the web.

## Install in one line

## What your agent can do

**Search the web in natural language**

Parallel Search lets agents declare what they need instead of constructing keyword queries. Results come back as ranked URLs with compressed, token-efficient excerpts, optimized for LLM context windows.

Every result includes dense, query-relevant excerpts. No raw HTML. No ads. No SEO noise.

****

**Extract clean content from any URL**

The Extract command converts any public URL into focused, AI-ready markdown. It handles JavaScript-rendered pages, CAPTCHAs, and PDFs automatically.

This pairs naturally with Search. An agent can find the right pages with search, then pull exactly what it needs with _Extract_.

****

**Run deep research on complex questions**

For questions that require synthesizing multiple sources, the Research command handles the entire workflow: searching, reading, cross-referencing, and producing cited reports.

Research outputs include citations, reasoning traces, and calibrated confidence scores through Parallel's Basis framework. Processor tiers range from _lite _(quick lookups, seconds) to _ultra_ (deep multi-source synthesis, up to 25 minutes).

**Enrich structured data at scale**

The Enrich command takes a CSV or JSON file and adds web-sourced fields to every row. Describe what you need in plain language, and Parallel handles the research for each record.

Enrichment scales from two-field lookups on the lite processor to 20+ field deep research on ultra. Pricing is per row, not per field, so adding more output columns doesn't change the cost.

## Get started

Install the CLI, grab an API key from platform.parallel.ai[platform.parallel.ai](https://platform.parallel.ai), and run your first search:

For structured agent integration, check out Agent Skills[Agent Skills](https://github.com/parallel-web/parallel-agent-skills) for pre-built skill definitions that teach agents how to use each command effectively.

Full documentation is at docs.parallel.ai[docs.parallel.ai](https://docs.parallel.ai/integrations/cli). The CLI source is open on GitHub[GitHub](https://github.com/parallel-web/parallel-web-tools).