cmd

README

desearch-cli

A fast, scriptable CLI for Desearch AI — a contextual AI search engine that aggregates results across web, Hacker News, Reddit, Wikipedia, YouTube, Twitter/X, and arXiv with AI-synthesized answers and citations.

Single binary, no Python or Node.js dependencies.

Install

Homebrew (recommended)

brew tap roboalchemist/tap
brew install desearch-cli

Binary

Download the latest release for your platform from the releases page, then:

# macOS (Intel or Apple Silicon)
curl -L https://github.com/roboalchemist/desearch-cli/releases/latest/download/desearch-cli_Darwin_arm64.tar.gz | tar xz
sudo mv desearch-cli /usr/local/bin/

# Linux
curl -L https://github.com/roboalchemist/desearch-cli/releases/latest/download/desearch-cli_Linux_x86_64.tar.gz | tar xz
sudo mv desearch-cli /usr/local/bin/

Build from source

go install github.com/roboalchemist/desearch-cli@latest

Setup

Get an API key at https://console.desearch.ai, then configure it:

desearch-cli config --api-key <YOUR_KEY>

That's it. The API key is stored in your config file at ~/.config/desearch-cli/config.toml.

Basic Usage

desearch-cli "golang best practices"

Commands

search (default)

desearch-cli [query] [flags]

Search across all or specified sources. This is the default command, so desearch-cli "query" and desearch-cli search "query" are equivalent.

completion

desearch-cli completion <query> [flags]

Stream an AI-generated summary without per-source search results.

config

desearch-cli config [flags]

Manage API key and default settings.

version

desearch-cli version

Show version information.

Search Flags

Flag	Type	Description
--query string	string	Search prompt (also the positional arg)
--tool strings	[]string	Sources to query: web, hackernews, reddit, wikipedia, youtube, twitter, arxiv (default: all)
--date-filter string	string	Predefined date range. One of: PAST_24_HOURS, PAST_2_DAYS, PAST_WEEK, PAST_2_WEEKS, PAST_MONTH, PAST_2_MONTHS, PAST_YEAR, PAST_2_YEARS
--start-date string	string	ISO8601 UTC start bound, e.g. 2026-01-01T00:00:00Z
--end-date string	string	ISO8601 UTC end bound
--streaming	bool	Stream results as they arrive
--result-type string	string	ONLY_LINKS or LINKS_WITH_FINAL_SUMMARY (default: LINKS_WITH_FINAL_SUMMARY)
--count int	int	Results per source, 10-200 (default: 10)
--system-message string	string	Override system prompt to influence AI behavior
--json	bool	Output raw JSON instead of formatted human-readable
--no-ai	bool	Skip AI completion/summary

Config Flags

Flag	Description
--api-key string	Set API key (get at https://console.desearch.ai)
--default-tool strings	Set default sources (can be specified multiple times)
--default-date-filter string	Set default date filter
--show	Display current configuration (alias: desearch-cli config show)
--clear	Reset configuration to defaults

Examples

Search with specific sources

desearch-cli "llm benchmarks" --tool web --tool hackernews
desearch-cli "rust vs go performance" --tool reddit --tool youtube

Date filtering

desearch-cli "AI news" --date-filter PAST_2_DAYS
desearch-cli "golang updates" --start-date 2026-01-01T00:00:00Z --end-date 2026-03-01T00:00:00Z

Limit results

desearch-cli "react patterns" --tool web --count 20

Custom system message

desearch-cli "explain quantum computing" --system-message "Summarize in simple terms for a non-technical audience"

Streaming output

desearch-cli "latest AI research" --streaming

AI completion only (no per-source results)

desearch-cli completion "what is bittensor"
desearch-cli completion "explain transformers" --system-message "Summarize in simple terms"

JSON output

desearch-cli "golangci-lint" --json | jq '.search[0].title'
desearch-cli "linux kernel news" --tool hackernews --json

Skip AI summary

desearch-cli "hacker news top posts" --no-ai

Set defaults

desearch-cli config --api-key sk-xxxx --default-tool web --default-date-filter PAST_WEEK

View current config

desearch-cli config show

Clear config

desearch-cli config clear

Configuration File

Config is stored at ~/.config/desearch-cli/config.toml (XDG spec).

api_key = "YOUR_KEY"
default_tools = ["web", "hackernews", "reddit", "wikipedia", "youtube", "twitter", "arxiv"]
default_date_filter = "PAST_24_HOURS"
default_count = 10

All fields are optional. Any field not set uses the API server defaults.

Output Format

By default, output is formatted with section headers per source:

=== WEB ===
[Article Title](https://example.com)
  Article snippet text

[Another Article](https://another.com)
  Another snippet

=== HACKERNEWS ===
[HN Post Title](https://news.ycombinator.com/item?id=123)
  Post snippet

=== REDDIT ===
[Reddit Post Title](https://reddit.com/r/...)
  Post snippet

=== TWITTER ===
@username
  Tweet text
  Link: https://x.com/...
  42 likes, 12 retweets, 5 replies

=== AI SUMMARY ===
AI-generated synthesis of the results...

Twitter results include engagement metrics (likes, retweets, replies, quotes, bookmarks) when available.

Use --json for raw JSON output.

License

MIT

Issue Tracker

Report bugs and request features at: https://github.com/roboalchemist/desearch-cli/issues

Documentation

Index

• func Execute() error
• func GetJSONOut() bool
• func RootCmd() *cobra.Command

Functions

func Execute

func Execute() error

Execute adds all child commands to the root command and sets flags appropriately.

func GetJSONOut

func GetJSONOut() bool

GetJSONOut returns whether --json flag was set.

func RootCmd

func RootCmd() *cobra.Command

RootCmd returns the root cobra command for use by gendocs.