NEXI for Agents

NEXI has three main command surfaces for agents:

• nexi - full agentic search with synthesized answers and citations
• nexi-search - direct search-provider execution without the agent loop
• nexi-fetch - direct fetch or extraction without the agent loop

Invocation

Pick the command that matches the job:

nexi --plain "your search query"
nexi-search --json "your search query"
nexi-fetch --json "https://example.com"

• Use nexi --plain when you want a final synthesized answer with citations.
• Use nexi-search --json when you want raw search results for scripts or agent post-processing.
• Use nexi-fetch --json when you want fetched page content or extraction payloads.

nexi --plain output is plain text with citations in [1], [2] format and a sources section at the end.

Effort Levels

Control agent search depth with -e; NEXI keeps the exact iteration and token budgets internal:

Example:

nexi --plain -e l "compare React vs Svelte performance 2024"

Output Format

[Answer text with inline citations like [1] and [2]...]

Sources:
[1] https://example.com/page1 - "Page Title"
[2] https://example.com/page2 - "Another Title"

Parse the sources section to extract URLs for further processing if needed.

Additional Flags

Config Lifecycle

If the config file does not exist yet, NEXI creates ~/.config/nexi/config.toml, tells you it is incomplete, and exits.

Useful commands:

nexi config
nexi onboard
nexi doctor

• nexi config opens the current config file
• nexi onboard guides a basic LLM plus search setup
• nexi doctor checks whether nexi, nexi-search, and nexi-fetch are ready

Prerequisites

NEXI requires configuration before use. See references/installation.md for:

• Install commands
• Required API keys and default fetch setup
• Config file location and format

If NEXI fails with config errors, read the installation reference to help set it up.