### Configuration File Example

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Example of a JSON configuration file for Brave Search CLI. Settings include API key, base URL, and timeout.

```json
{
  "api_key": "BSA...",
  "base_url": "https://api.search.brave.com",
  "timeout": 30
}
```

--------------------------------

### Install Brave Search CLI on Windows (PowerShell)

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Use this PowerShell command to install the bx CLI on Windows systems.

```powershell
powershell -ExecutionPolicy Bypass -c "irm https://raw.githubusercontent.com/brave/brave-search-cli/main/scripts/install.ps1 | iex"
```

--------------------------------

### Load Custom Config File

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Example of using the `--config` flag to load a specific configuration file for different environments.

```bash
bx --config ~/.config/brave-search/dev.json web "test"
```

--------------------------------

### Install Brave Search CLI on macOS/Linux

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Use this command to install the bx CLI on macOS and Linux systems via curl.

```bash
curl -fsSL https://raw.githubusercontent.com/brave/brave-search-cli/main/scripts/install.sh | sh
```

--------------------------------

### RAG Grounding with Token Budget

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Example of using the 'context' command for RAG/LLM grounding, specifying maximum tokens and a strict threshold.

```bash
# RAG grounding with token budget
bx context "Python TypeError cannot unpack non-iterable NoneType" --max-tokens 4096 --threshold strict
```

--------------------------------

### Default Command Usage

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Example of running the CLI without specifying a subcommand, which defaults to the 'context' command.

```bash
# Default subcommand is context
bx "tokio spawn async task example"
```

--------------------------------

### Web Search with Site Scoping and jq

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Example of performing a web search using the 'web' command, with site-specific scoping and piping results to 'jq' to extract URLs.

```bash
# Web search with site: scoping
bx web "site:docs.rs axum middleware" | jq '.web.results[].url'
```

--------------------------------

### Image Search and Thumbnail Extraction

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Example of using the 'images' command for image search and extracting thumbnail sources using 'jq'.

```bash
# Image search
bx images "system architecture diagram microservices" | jq '.results[].thumbnail.src'
```

--------------------------------

### AI Answer Streaming

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Example of using the 'answers' command to get AI-generated answers, with streaming enabled by default.

```bash
# AI answer (streaming)
bx answers "explain Rust lifetimes with examples"
```

--------------------------------

### Autocomplete Suggestions with Count and jq

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Example of using the 'suggest' command for query autocomplete, specifying the number of suggestions and extracting queries with 'jq'.

```bash
# Autocomplete
bx suggest "how to implement" --count 10 | jq '.results[].query'
```

--------------------------------

### Get Autocomplete/Query Suggestions

Source: https://context7.com/brave/brave-search-cli/llms.txt

Use `bx suggest` to get query suggestions for autocomplete. The `--rich` flag provides metadata, and `--lang` and `--country` can filter suggestions.

```bash
bx suggest "how to implement" --count 10 | jq '.results[].query'
```

```bash
bx suggest "Rust error handling" --rich | jq .
```

```bash
bx suggest "machine learning" --lang en --country US --count 5
```

--------------------------------

### Agent Workflow: Debugging an Error

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Example of using `bx context` to debug a specific Python error, with a token limit.

```bash
# Agent encounters: "TypeError: cannot unpack non-iterable NoneType"
bx "Python TypeError cannot unpack non-iterable NoneType" --max-tokens 4096
```

--------------------------------

### Location-Aware Web Search

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Example of performing a web search with location awareness using latitude, longitude, and city parameters.

```bash
# Location-aware web search
bx web "restaurants near me" --lat 37.7749 --long -122.4194 --city "San Francisco"
```

--------------------------------

### AI Answer Non-Streaming with jq

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Example of using the 'answers' command to get AI-generated answers without streaming, piping the output to 'jq' for JSON parsing.

```bash
# AI answer (non-streaming)
bx answers "compare SQLx and Diesel for Rust database access" --no-stream | jq .
```

--------------------------------

### Local Proxy / Sandboxed Agents Setup

Source: https://context7.com/brave/brave-search-cli/llms.txt

Configure `bx` to point to a local reverse proxy for injecting API credentials outside the sandbox. Only loopback addresses are accepted.

```bash
# Point bx at a local reverse proxy
# bx --proxy http://127.0.0.1:8080 ...
```

--------------------------------

### Focused Search with Goggles (Custom Re-ranking)

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Example of using Goggles to boost official docs, prefer specific paths, and suppress SEO spam for a search query.

```bash
# Boost official docs, prefer /docs/ paths, suppress SEO spam
bx "Python asyncio gather vs wait" \
  --goggles '$boost=3,site=docs.python.org
/docs/$boost=3
/blog/$downrank=2
$discard,site=geeksforgeeks.org
$discard,site=w3schools.com' --max-tokens 4096
```

--------------------------------

### Local Places Search with jq

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Example of searching for local places using the 'places' command, specifying location and query, and extracting titles with 'jq'.

```bash
# Local places
bx places --location "San Francisco CA US" -q "coffee" | jq '.results[].title'
```

--------------------------------

### Agent Workflow: Corrective RAG Loop

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Demonstrates a RAG loop: starting with a broad search, narrowing down, and then asking for a synthesized answer.

```bash
# 1. Broad search
bx "axum middleware authentication" --max-tokens 4096
```

```bash
# 2. Too general? Narrow down
bx "axum middleware tower layer authentication example" --threshold strict --max-tokens 4096
```

```bash
# 3. Still need synthesis? Ask for an answer
bx answers "how to implement JWT auth middleware in axum" --enable-research
```

--------------------------------

### Research Mode for AI Answers

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Example of enabling research mode with the 'answers' command for more thorough AI-generated responses.

```bash
# Research mode (multi-step, thorough)
bx answers "what changed in React 19 vs React 18?" --enable-research
```

--------------------------------

### Boost Official Sources for Python Errors with Goggles

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Apply Goggles rules to boost or discard specific domains for a given query. This example prioritizes official Python documentation and GitHub.

```bash
bx "TypeError cannot unpack non-iterable NoneType" \
  --goggles '$boost=3,site=docs.python.org
$boost=2,site=github.com
$discard,site=geeksforgeeks.org
$discard,site=w3schools.com'
```

--------------------------------

### Direct AI Answer with Streaming

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Get a direct AI-generated answer to a query. This command streams the response by default.

```bash
bx answers "explain Rust lifetimes with examples"
```

--------------------------------

### `bx web` Response Shape

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Example JSON structure for the `bx web` command, including web, news, videos, and discussions results.

```json
{
  "web": { "results": [{"title": "...", "url": "...", "description": "..."}] },
  "news": { "results": [...] },
  "videos": { "results": [...] },
  "discussions": { "results": [...] }
}
```

--------------------------------

### News Search with Freshness Filter and jq

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Example of searching for news articles using the 'news' command, filtering by freshness (past day), and extracting titles with 'jq'.

```bash
# News from the past day
bx news "npm security advisory" --freshness pd | jq '.results[].title'
```

--------------------------------

### `bx answers` Response Shape (Single Response)

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Example JSON structure for a non-streamed response from the `bx answers` command.

```json
{"choices": [{"message": {"content": "Rust lifetimes ensure references..."}}]}
```

--------------------------------

### `bx context` Response Shape

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Example JSON structure for the `bx context` command, showing grounding results with extracted content.

```json
{
  "grounding": {
    "generic": [
      { "url": "...", "title": "...", "snippets": ["extracted content...", "..."] }
    ]
  }
}
```

--------------------------------

### `bx answers` Response Shape (Streaming)

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Example JSON structure for a single chunk of a streamed response from the `bx answers` command.

```json
{"choices": [{"delta": {"content": "Rust lifetimes ensure references..."}}]}
```

--------------------------------

### Web Search with Result Type Filtering

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Example of filtering web search results to include only specific types like 'web' and 'discussions'.

```bash
# Filter result types
bx web "rust" --result-filter "web,discussions"
```

--------------------------------

### Error Handling Output Example

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Errors are printed to stderr with a human-readable summary, recovery hints, and the full JSON error body. Exit codes differentiate error types.

```text
error: rate limited (429) — Request rate limit exceeded for plan.
hint: retry after a short delay, or upgrade plan for higher rate limits
{"type":"ErrorResponse","error":{"code":"RATE_LIMITED","status":429,...}}
```

--------------------------------

### Override API Path with --endpoint

Source: https://context7.com/brave/brave-search-cli/llms.txt

Use the `--endpoint` flag to override the default API path, useful for accessing beta or internal endpoints. The path must start with `/` and contain only allowed characters.

```bash
# Use a beta API version
bx web "test" --endpoint /res/v2/web/search
```

```bash
# Combined with --extra for full control
bx web "rust" \
  --endpoint /res/v1/web/search \
  --extra result_filter=web \
  --extra count=5
```

--------------------------------

### Override Default API Path with --endpoint

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Use the --endpoint flag to override the default API path, useful for internal or beta Brave endpoints. The path must start with '/'.

```bash
bx web "test" --endpoint /res/v2/web/search
```

--------------------------------

### Interpret Brave Search CLI Exit Codes

Source: https://context7.com/brave/brave-search-cli/llms.txt

Understand the meaning of different exit codes returned by Brave Search CLI commands for effective error handling in scripts. Includes a bash script example for checking exit codes.

```bash
# Exit codes:
# 0 = Success
# 1 = Client error (bad request / 4xx general)
# 2 = Usage error (bad CLI flags — clap)
# 3 = Auth/permission error (401/403)
# 4 = Rate limited (429)
# 5 = Server/network error (5xx, timeout)

# Check exit code in a script
if bx context "query" --max-tokens 2048 > output.json; then
  echo "Success"
  jq '.grounding.generic[0].snippets[0]' output.json
else
  code=$?
  case $code in
    3) echo "Auth error — run: bx config show-key" ;;
    4) echo "Rate limited — retrying in 5s..."; sleep 5 ;; 
    5) echo "Server error — check network" ;; 
    *) echo "Error (code $code)" ;; 
  esac
fi
```

```text
# Sample stderr error output:
# error: rate limited (429) — Request rate limit exceeded for plan.
# hint: retry after a short delay, or upgrade plan for higher rate limits
# {"type":"ErrorResponse","error":{"code":"RATE_LIMITED","status":429,...}}
```

--------------------------------

### Pass Additional API Parameters with --extra

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Use the --extra flag to pass additional API parameters not covered by named flags. Values are auto-typed for POST endpoints (numbers, booleans, strings) and always strings for GET endpoints. Be aware of potential key collisions with existing flags.

```bash
bx web "rust" --extra custom_param=value
```

```bash
bx web "rust" --extra count=5            # overrides --count (with warning)
```

```bash
bx images "sunset" --extra safesearch=off
```

--------------------------------

### Build Brave Search CLI from Source

Source: https://context7.com/brave/brave-search-cli/llms.txt

Instructions for building the Brave Search CLI from source using Cargo, including debug and release builds, and cross-compilation using Docker. Also covers running unit and end-to-end tests.

```bash
# Prerequisites: Rust toolchain (see rust-toolchain.toml for pinned version)

# Debug build
cargo build

# Optimized release build (~2 MB, fully stripped)
cargo build --release
./target/release/bx --version

# Static cross-compiled binaries via Docker
make linux-amd64      # → dist/bx-<version>-linux-amd64
make linux-arm64      # → dist/bx-<version>-linux-arm64
make darwin-arm64     # → dist/bx-<version>-darwin-arm64
make windows-amd64    # → dist/bx-<version>-windows-amd64.exe
make dist-all         # → all platforms

# Run unit tests (no API key required)
make test-unit
cargo test

# Run end-to-end API smoke tests (requires BRAVE_SEARCH_API_KEY and jq)
make test
```

--------------------------------

### Build Static Binaries with Docker

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Produce fully static, cross-compiled binaries using Docker, GNU Make, and a POSIX shell. Requires Docker. On Windows, use Git Bash or WSL. The Makefile uses POSIX utilities.

```bash
make linux-amd64     # → dist/bx-<version>-linux-amd64
make linux-arm64     # → dist/bx-<version>-linux-arm64
make darwin-arm64    # → dist/bx-<version>-darwin-arm64
make windows-amd64   # → dist/bx-<version>-windows-amd64.exe
make windows-arm64   # → dist/bx-<version>-windows-arm64.exe
make dist-all        # → all of the above
```

--------------------------------

### Configure API Key and Basic Usage

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Set your Brave Search API key and learn basic command usage, including shorthand for context queries and accessing help.

```bash
bx config set-key YOUR_API_KEY    # get a key at https://api-dashboard.search.brave.com
```

```bash
bx "your search query"            # shorthand for: bx context "your search query"
```

```bash
bx --help                         # see all commands; bx <command> --help for flags
```

```bash
# To search for a word matching a subcommand name: bx -- web  or  bx context "web"
```

--------------------------------

### Set Up IPv6 Loopback

Source: https://context7.com/brave/brave-search-cli/llms.txt

Configure the Brave Search CLI to use a local IPv6 loopback address. This is for environments requiring IPv6 connectivity.

```bash
BRAVE_SEARCH_BASE_URL=http://[::1]:8080 \
BRAVE_SEARCH_API_KEY=unused \
bx web "rust async"
```

--------------------------------

### Agent Workflow: Evaluating a Dependency

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Check dependency maintenance and security issues using `bx context` and `bx news`.

```bash
# Is this crate maintained? Any security issues?
bx context "reqwest crate security issues maintained 2026" --threshold strict
```

```bash
bx news "reqwest Rust crate" --freshness pm
```

--------------------------------

### Error Debugging with Domain Boosting

Source: https://context7.com/brave/brave-search-cli/llms.txt

Demonstrates using `bx context` with domain boosting (`--goggles`) to focus search results on specific sites for effective error debugging. Helps in finding solutions for complex issues.

```bash
# Error debugging with domain boosting
bx context "TypeError cannot unpack non-iterable NoneType Python" \
  --goggles '$boost=3,site=docs.python.org
$boost=2,site=stackoverflow.com
$discard,site=geeksforgeeks.org' \
  --max-tokens 4096
```

--------------------------------

### Version/Release Checking Beyond Training Cutoff

Source: https://context7.com/brave/brave-search-cli/llms.txt

Shows how to use `bx context` and `bx news` to retrieve information about the latest releases or changelogs for software, especially when the information is newer than the CLI's training data.

```bash
# Version/release checking (beyond training cutoff)
bx context "tokio 1.x latest release changelog" --max-tokens 2048
bx news "tokio rust async release" --freshness pm
```

--------------------------------

### Local Proxy Configuration

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Demonstrates setting the `BRAVE_SEARCH_BASE_URL` environment variable to point the CLI to a local proxy. A placeholder API key is required when using a proxy.

```bash
# Point bx at a local proxy on port 8080 with path prefix /brave
BRAVE_SEARCH_BASE_URL=http://127.0.0.1:8080/brave bx "your query"

# Or in config.json
# { "base_url": "http://127.0.0.1:8080/brave" }
```

--------------------------------

### Dependency Evaluation Pipeline with Brave Search CLI

Source: https://context7.com/brave/brave-search-cli/llms.txt

Shows how to use `bx context` and `bx news` to evaluate dependencies, checking for security issues or recent changes. Useful for maintaining project health.

```bash
# Dependency evaluation pipeline
bx context "reqwest crate security issues maintained 2025" --threshold strict --max-tokens 2048
bx news "reqwest Rust crate" --freshness pm | jq '.results[].title'
```

--------------------------------

### Focus on Rust Ecosystem Sources with Goggles

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Boost official Rust documentation and source repositories like docs.rs, crates.io, and GitHub for crate research.

```bash
bx context "serde custom deserializer" \
  --goggles '$boost=5,site=docs.rs
$boost=5,site=crates.io
$boost=3,site=github.com' --max-tokens 4096
```

--------------------------------

### Set Up IPv4 Loopback Proxy with Path Prefix

Source: https://context7.com/brave/brave-search-cli/llms.txt

Configure the Brave Search CLI to use a local IPv4 loopback proxy with a specific path prefix. This is useful for testing or custom integrations.

```bash
BRAVE_SEARCH_BASE_URL=http://127.0.0.1:8080/brave \
BRAVE_SEARCH_API_KEY=unused \
bx context "axum middleware" --max-tokens 4096
```

--------------------------------

### Create and Use Goggles Rules from a File

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Define Goggles rules in a file and reference it using the '--goggles @file' syntax. This is ideal for agents to manage complex ranking preferences.

```bash
cat > /tmp/rust.goggle << 'EOF'
$boost=5,site=docs.rs
$boost=5,site=crates.io
$boost=3,site=github.com
/blog/$downrank=3
$discard,site=w3schools.com
$discard,site=geeksforgeeks.org
EOF

bx context "axum middleware tower" --goggles @/tmp/rust.goggle --max-tokens 4096
bx context "serde custom deserializer" --goggles @/tmp/rust.goggle --max-tokens 4096
```

--------------------------------

### Prioritize Docs Paths with Goggles

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Use Goggles to boost specific URL path patterns, such as '/docs/', over others like '/blog/'. This helps in finding more relevant documentation.

```bash
bx context "axum middleware tower" \
  --goggles '/docs/$boost=5
/api/$boost=3
/blog/$downrank=3' --max-tokens 4096
```

--------------------------------

### Breaking Change Lookup with Brave Search CLI

Source: https://context7.com/brave/brave-search-cli/llms.txt

Illustrates how to use `bx context` and `bx news` to find breaking changes in software versions before upgrading. Helps in planning migration strategies.

```bash
# Breaking change lookup before upgrading
bx context "Next.js 15 breaking changes migration guide" --max-tokens 8192
bx news "Next.js 15 release" --freshness pm
```

--------------------------------

### Corrective RAG Loop with Brave Search CLI

Source: https://context7.com/brave/brave-search-cli/llms.txt

Demonstrates using `bx context` for a corrective RAG loop, showing how to refine queries with `--threshold` and `--max-tokens` for better LLM input. Escalates to `bx answers` for synthesis tasks.

```bash
# Corrective RAG loop
bx context "axum middleware authentication" --max-tokens 4096
# Too general? Narrow down with --threshold strict
bx context "axum middleware tower layer JWT authentication example" --threshold strict --max-tokens 4096
# Need synthesis? Escalate to answers
bx answers "how to implement JWT auth middleware in axum" --enable-research
```

--------------------------------

### Build Brave Search CLI with Cargo

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Build the Brave Search CLI from source using Cargo. Use 'cargo build' for a debug build and 'cargo build --release' for a smaller release build.

```bash
# Debug build
cargo build

# Release build (~2 MB)
cargo build --release
```

--------------------------------

### Show Brave Search CLI Config File Path

Source: https://context7.com/brave/brave-search-cli/llms.txt

Prints the path to the Brave Search CLI configuration file. This helps in locating and manually editing the configuration.

```bash
bx config path
```

--------------------------------

### Agent Workflow: Checking for Breaking Changes

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Use `bx context` and `bx news` to check for breaking changes before upgrading a dependency.

```bash
bx context "Next.js 15 breaking changes migration guide" --max-tokens 8192
```

```bash
bx news "Next.js 15 release" --freshness pm
```

--------------------------------

### Show Full Brave Search CLI Configuration

Source: https://context7.com/brave/brave-search-cli/llms.txt

Displays the complete Brave Search CLI configuration, including API key (masked), base URL, and timeout settings.

```bash
bx config show
```

--------------------------------

### Using Faster Model in Brave CLI

Source: https://context7.com/brave/brave-search-cli/llms.txt

Switches to a faster, potentially less thorough, model for quicker responses.

```bash
bx answers "Hello" --model brave
```

--------------------------------

### Basic RAG Grounding with bx context

Source: https://context7.com/brave/brave-search-cli/llms.txt

Performs basic RAG grounding using the `bx context` command. This is a shorthand for `bx context "query"`.

```bash
bx "Python TypeError cannot unpack non-iterable NoneType"
```

--------------------------------

### Use Hosted Goggles URL for Search Ranking

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Reference a Goggles configuration file hosted online (e.g., on GitHub) using its URL with the '--goggles' flag. This enables sharing and reusing ranking configurations.

```bash
bx web "distributed systems" \
  --goggles 'https://raw.githubusercontent.com/brave/goggles-quickstart/main/goggles/tech_blogs.goggle'
```

--------------------------------

### Citations and Entities in Brave CLI (Streaming)

Source: https://context7.com/brave/brave-search-cli/llms.txt

Enables streaming of results with citations and identified entities. Useful for real-time information retrieval.

```bash
bx answers "who invented the internet?" --enable-citations --enable-entities
```

--------------------------------

### Configure Brave Search API Key Directly

Source: https://context7.com/brave/brave-search-cli/llms.txt

Sets the Brave Search API key directly via a command-line argument. Replace 'BSA_YOUR_KEY_HERE' with your actual API key.

```bash
bx config set-key BSA_YOUR_KEY_HERE
```

--------------------------------

### Stdin Mode for Brave CLI Answers

Source: https://context7.com/brave/brave-search-cli/llms.txt

Processes a full OpenAI-compatible JSON body piped via stdin. Use '-' as the argument for answers.

```bash
echo '{"messages":[{"role":"user","content":"review this code for security issues"}],"model":"brave-pro","stream":false}' \
  | bx answers -
```

--------------------------------

### RAG Grounding with Token Budget

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Use the `context` command for RAG grounding, specifying a maximum token limit for the output.

```bash
bx context "Python TypeError cannot unpack non-iterable NoneType" --max-tokens 4096
```

--------------------------------

### Configure Brave Search API Key Interactively

Source: https://context7.com/brave/brave-search-cli/llms.txt

Sets the Brave Search API key by prompting the user interactively. This method avoids storing the key in shell history.

```bash
bx config set-key
```

--------------------------------

### Include and Exclude Sites with Brave Search CLI

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Use --include-site and --exclude-site flags to filter search results by domain.

```bash
bx "rust axum" --include-site docs.rs --include-site github.com
```

```bash
bx web "rust tutorial" --exclude-site w3schools.com --exclude-site medium.com
```

--------------------------------

### RAG Grounding with Token Budget Control

Source: https://context7.com/brave/brave-search-cli/llms.txt

Executes RAG grounding with specific controls for maximum tokens, tokens per URL, number of URLs, and a strictness threshold.

```bash
bx context "axum middleware tower layer authentication" \
  --max-tokens 4096 \
  --max-tokens-per-url 1024 \
  --max-urls 5 \
  --threshold strict
```

--------------------------------

### Extract Snippets with jq

Source: https://context7.com/brave/brave-search-cli/llms.txt

Demonstrates how to extract specific fields (URL and text snippets) from the RAG grounding output using `jq`.

```bash
bx context "Rust ownership model" --max-tokens 2048 | \
  jq '.grounding.generic[] | {url, text: .snippets[]}'
```

--------------------------------

### Raw Web Search with Scoping

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Perform a raw web search, allowing for site scoping and result count specification.

```bash
bx web "site:docs.rs axum middleware" --count 5
```

--------------------------------

### Video Search (Past Month) with jq Extraction

Source: https://context7.com/brave/brave-search-cli/llms.txt

Searches for tutorial videos from the past month and extracts title, URL, and duration using jq.

```bash
bx videos "Rust async await tutorial" \
  --freshness pm \
  --count 5 | \
  jq '.results[] | {title, url, duration: .video.duration}'
```

--------------------------------

### Run Brave Search CLI Tests

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Execute tests for the Brave Search CLI. Use 'make test' for end-to-end API smoke tests (requires API key) or 'make test-unit' for unit tests (no network required).

```bash
# End-to-end API smoke suite (requires bash + jq + BRAVE_SEARCH_API_KEY/BRAVE_API_KEY)
make test

# Unit tests (no network/API key required)
make test-unit
```

--------------------------------

### Configure Brave Search API Key via Environment Variable

Source: https://context7.com/brave/brave-search-cli/llms.txt

Sets the Brave Search API key using the `BRAVE_SEARCH_API_KEY` environment variable. This is a common method for CI/CD or server environments.

```bash
export BRAVE_SEARCH_API_KEY=BSA_YOUR_KEY_HERE
```

--------------------------------

### Piping Request File to Brave CLI

Source: https://context7.com/brave/brave-search-cli/llms.txt

Reads a JSON request from a file, pipes it to the Brave CLI for processing, and then uses jq to extract the content.

```bash
cat request.json | bx answers - | jq '.choices[0].message.content'
```

--------------------------------

### Local Place Search by Location String

Source: https://context7.com/brave/brave-search-cli/llms.txt

Searches for local places (e.g., coffee shops) using a location string and extracts name and address using jq.

```bash
bx places "coffee" --location "San Francisco CA US" | \
  jq '.results[] | {name: .title, address: .postal_address}'
```

--------------------------------

### Image Search with jq Extraction

Source: https://context7.com/brave/brave-search-cli/llms.txt

Searches for images and extracts the title and thumbnail source URL using jq.

```bash
bx images "microservices architecture diagram" --count 10 | \
  jq '.results[] | {title, thumbnail: .thumbnail.src}'
```

--------------------------------

### RAG Grounding with Goggles DSL for Re-ranking

Source: https://context7.com/brave/brave-search-cli/llms.txt

Applies advanced re-ranking to RAG grounding results using Brave's Goggles DSL. This allows fine-grained control over search result boosting and discarding.

```bash
bx context "Python asyncio gather vs wait" \
  --goggles '$boost=3,site=docs.python.org
/docs/$boost=3
/blog/$downrank=2
$discard,site=geeksforgeeks.org
$discard,site=w3schools.com' \
  --max-tokens 4096
```

--------------------------------

### Custom Search Re-Ranking with Goggles

Source: https://context7.com/brave/brave-search-cli/llms.txt

Brave Goggles allow custom re-ranking rules (boost, downrank, discard) by domain or URL path. Rules can be inline, multiline via \n, from a file, or from stdin. Supported on `context`, `web`, and `news`.

```bash
# Inline rules (single quotes required to prevent shell expansion of $)
bx context "serde custom deserializer" \
  --goggles '$boost=5,site=docs.rs
$boost=5,site=crates.io
$boost=3,site=github.com
/blog/$downrank=3
$discard,site=w3schools.com' \
  --max-tokens 4096
```

```bash
# Multiline via \n escape in single-line form
bx web "tokio runtime" \
  --goggles '$boost=5,site=docs.rs\n$discard,site=medium.com'
```

```bash
# From a file (ideal for agents — generate once, reuse across queries)
cat > /tmp/rust.goggle << 'EOF'
$boost=5,site=docs.rs
$boost=5,site=crates.io
$boost=3,site=github.com
/blog/$downrank=3
$discard,site=w3schools.com
$discard,site=geeksforgeeks.org
EOF

bx context "axum middleware tower" --goggles @/tmp/rust.goggle --max-tokens 4096
bx context "serde custom deserializer" --goggles @/tmp/rust.goggle --max-tokens 4096
```

```bash
# From stdin
echo '$boost=5,site=docs.rs
$boost=3,site=github.com' | bx web "tokio runtime" --goggles @-
```

```bash
# Hosted goggle URL (community-maintained)
bx web "distributed systems" \
  --goggles 'https://raw.githubusercontent.com/brave/goggles-quickstart/main/goggles/tech_blogs.goggle'
```

```bash
# --include-site / --exclude-site (convenience shortcuts for common cases)
bx context "axum middleware" \
  --include-site docs.rs \
  --include-site github.com \
  --max-tokens 4096
```

```bash
bx web "rust tutorial" --exclude-site w3schools.com --exclude-site medium.com
```

--------------------------------

### Token Budget Control

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Control the token budget for a search query by setting maximum tokens, maximum tokens per URL, and maximum URLs.

```bash
bx context "topic" --max-tokens 4096 --max-tokens-per-url 1024 --max-urls 5
```

--------------------------------

### Pipe Goggles Rules from Stdin

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Provide Goggles rules dynamically by piping them from another command using '--goggles @-'. This allows for flexible rule generation.

```bash
echo '$boost=5,site=docs.rs
$boost=3,site=github.com' | bx web "tokio runtime" --goggles @-
```

--------------------------------

### Basic Web Search with jq Extraction

Source: https://context7.com/brave/brave-search-cli/llms.txt

Performs a basic web search and uses jq to extract the URLs from the web results.

```bash
bx web "site:docs.rs axum middleware" | jq '.web.results[].url'
```

--------------------------------

### Browse Places by Location

Source: https://context7.com/brave/brave-search-cli/llms.txt

Use the `bx places` command to browse locations without a specific query. Pipe output to `jq` for structured data extraction.

```bash
bx places --location "NYC" --count 5 | jq '.results[].title'
```

```bash
bx places "museum" --location "Paris, France" --radius "2000" --count 10
```

```bash
bx places "gas station" \
  --location "London UK" \
  --units metric \
  --ui-lang en-GB | \
  jq '.results[] | {name: .title, contact}'
```

--------------------------------

### Pagination in Web Search

Source: https://context7.com/brave/brave-search-cli/llms.txt

Retrieves results with specified count and offset for pagination. Useful for fetching data in chunks.

```bash
bx web "rust async runtime" --count 10 --offset 20
```

--------------------------------

### RAG Grounding with Domain Allowlist

Source: https://context7.com/brave/brave-search-cli/llms.txt

Restricts RAG grounding results to specific domains using `--include-site`. Useful for focusing on trusted or relevant sources.

```bash
bx context "reqwest crate usage examples" \
  --include-site docs.rs \
  --include-site github.com \
  --max-tokens 4096
```

--------------------------------

### Image Search by Language and Region

Source: https://context7.com/brave/brave-search-cli/llms.txt

Searches for images in a specific language and country, extracting URLs using jq.

```bash
bx images "traditional Japanese architecture" \
  --country JP \
  --search-lang ja \
  --count 15 | \
  jq '.results[].url'
```

--------------------------------

### Video Search with Operators

Source: https://context7.com/brave/brave-search-cli/llms.txt

Uses search operators to refine video searches, focusing on titles and retrieving up to 10 results.

```bash
bx videos "intitle:\"Rust tutorial\" beginner" --operators --count 10
```

--------------------------------

### Local Place Search by Coordinates

Source: https://context7.com/brave/brave-search-cli/llms.txt

Searches for local places (e.g., pizza restaurants) using geographic coordinates and limits results to 10.

```bash
bx places "pizza" --latitude 37.7749 --longitude -122.4194 --count 10
```

--------------------------------

### Enabling Search Operators in Web Search

Source: https://context7.com/brave/brave-search-cli/llms.txt

Activates the use of search operators like 'site:' and 'intitle:' for more precise web searches.

```bash
bx web "site:github.com rust-lang/rust issue borrow checker" --operators
```

--------------------------------

### Web Search with Extra Snippets

Source: https://context7.com/brave/brave-search-cli/llms.txt

Retrieves extra snippets along with web search results and extracts them using jq.

```bash
bx web "axum routing" --extra-snippets | jq '.web.results[0].extra_snippets'
```

--------------------------------

### Use Brave Answers Endpoint - Simple Mode

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Interact with the Brave Answers endpoint by passing a question directly. Supports streaming responses by default, single JSON responses with --no-stream, deep research with --enable-research, and faster models with --model.

```bash
bx answers "explain Rust lifetimes with examples"         # streaming (default)
```

```bash
bx answers "compare SQLx and Diesel" --no-stream          # single JSON response
```

```bash
bx answers "what changed in React 19?" --enable-research  # deep research
```

```bash
bx answers "Hello" --model brave                          # faster model
```

--------------------------------

### Location-Aware RAG Grounding

Source: https://context7.com/brave/brave-search-cli/llms.txt

Performs RAG grounding with location-specific context by providing latitude, longitude, and city information.

```bash
bx context "coffee shops open now" \
  --lat 37.7749 --long -122.4194 \
  --city "San Francisco" \
  --max-tokens 2048
```

--------------------------------

### RAG Grounding with Domain Exclusions

Source: https://context7.com/brave/brave-search-cli/llms.txt

Excludes specific domains from RAG grounding results using `--exclude-site`. Helps to filter out noisy or irrelevant websites.

```bash
bx context "Python asyncio tutorial" \
  --exclude-site w3schools.com \
  --exclude-site medium.com
```

--------------------------------

### Pass Arbitrary API Parameters with --extra

Source: https://context7.com/brave/brave-search-cli/llms.txt

The `--extra` flag allows passing arbitrary API parameters not covered by named flags. Values are auto-typed for POST endpoints. It warns on collisions with existing flags.

```bash
# Pass a custom parameter
bx web "rust" --extra custom_param=value
```

```bash
# Override count (warns: "--extra 'count=5' overrides existing parameter")
bx web "rust" --extra count=5
```

```bash
# Multiple extras
bx web "rust" --extra safesearch=off --extra count=10
```

```bash
# Integer, float, and boolean auto-typing in POST bodies
bx context "rust" --extra maximum_number_of_tokens=8192  # becomes JSON integer
bx context "rust" --extra enable_local=true              # becomes JSON boolean
```

```bash
# Images with extra parameter (GET endpoint — values are always strings)
bx images "sunset" --extra safesearch=off
```

--------------------------------

### Video Search with jq Extraction

Source: https://context7.com/brave/brave-search-cli/llms.txt

Performs a general video search and extracts the URLs of the results using jq.

```bash
bx videos "distributed systems design" --count 10 | \
  jq '.results[].url'
```

--------------------------------

### Research Mode with Fine-Grained Controls

Source: https://context7.com/brave/brave-search-cli/llms.txt

Configures deep research mode with specific limits on the number of queries and time spent, and adjusts search context size.

```bash
bx answers "explain Kubernetes networking" \
  --enable-research \
  --research-max-queries 5 \
  --research-max-seconds 60 \
  --search-context-size high
```

--------------------------------

### Show Masked Brave Search API Key

Source: https://context7.com/brave/brave-search-cli/llms.txt

Displays the configured Brave Search API key, masked for security. Useful for verifying the key is set without revealing it.

```bash
bx config show-key
```

--------------------------------

### Use Brave Answers Endpoint - Stdin Mode

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Read a full JSON request body from stdin for the Answers endpoint by passing '-'. This allows for complex queries or piping data from other commands.

```bash
echo '{"messages":[{"role":"user","content":"review this code for security issues"}],"model":"brave-pro"}' | bx answers -
```

```bash
cat request.json | bx answers -
```

--------------------------------

### Downrank Blog Spam in News Results with Goggles

Source: https://github.com/brave/brave-search-cli/blob/main/README.md

Use Goggles to downrank specific domains, like medium.com, in news search results to reduce noise from less authoritative sources.

```bash
bx news "npm security advisory" --freshness pd \
  --goggles '$downrank=5,site=medium.com'
```

--------------------------------

### Deep Research Mode in Brave CLI

Source: https://context7.com/brave/brave-search-cli/llms.txt

Enables multi-step, thorough research for complex queries. May take longer to complete.

```bash
bx answers "what changed in React 19 vs React 18?" --enable-research
```

--------------------------------

### News Search with Goggles and jq Extraction

Source: https://context7.com/brave/brave-search-cli/llms.txt

Filters news by past day, applies Goggles to downrank specific sites, and extracts title, URL, and age using jq.

```bash
bx news "npm security advisory" \
  --freshness pd \
  --goggles '$downrank=5,site=medium.com' | \
  jq '.results[] | {title, url, age}'
```

--------------------------------

### Date Range Freshness Filter in Web Search

Source: https://context7.com/brave/brave-search-cli/llms.txt

Specifies a custom date range for freshness filtering in web searches.

```bash
bx web "Rust 2024 edition" --freshness "2024-01-01to2024-12-31"
```

--------------------------------

### Location-Aware Web Search

Source: https://context7.com/brave/brave-search-cli/llms.txt

Performs a web search that is aware of the user's location, specified by coordinates, city, and country.

```bash
bx web "restaurants near me" \
  --lat 37.7749 --long -122.4194 \
  --city "San Francisco" \
  --loc-country US
```

--------------------------------

### Goggles for Web Search Ranking Control

Source: https://context7.com/brave/brave-search-cli/llms.txt

Uses Goggles to boost or discard specific sites, influencing search result ranking. Limits results to 5.

```bash
bx web "Python request session" \
  --goggles '$boost=5,site=docs.python-requests.org
$boost=3,site=github.com
$discard,site=geeksforgeeks.org' \
  --count 5
```

--------------------------------

### Limiting Completion Tokens in Brave CLI

Source: https://context7.com/brave/brave-search-cli/llms.txt

Restricts the maximum number of tokens in the completion response to control output length. Disables streaming.

```bash
bx answers "summarize the CAP theorem" --max-completion-tokens 256 --no-stream
```

--------------------------------

### Image Search with Safe Search

Source: https://context7.com/brave/brave-search-cli/llms.txt

Performs an image search with a strict safe search filter, retrieving up to 20 results.

```bash
bx images "chemistry lab equipment" --safesearch strict --count 20
```

--------------------------------

### Recent News Search (Past Month)

Source: https://context7.com/brave/brave-search-cli/llms.txt

Fetches recent news articles from the past month, limiting the results to 10.

```bash
bx news "OpenAI API updates" --freshness pm --count 10
```

--------------------------------

### News Search (Past Week) with jq Extraction

Source: https://context7.com/brave/brave-search-cli/llms.txt

Searches for news articles from the past week and extracts their titles using jq.

```bash
bx news "Rust language" --freshness pw | jq '.results[].title'
```

--------------------------------

### Freshness Filter (Past Day) in Web Search

Source: https://context7.com/brave/brave-search-cli/llms.txt

Filters web search results to include only those from the past day. Extracts URLs using jq.

```bash
bx web "npm security advisory" --freshness pd | jq '.web.results[].url'
```

--------------------------------

### Spell Check and Correct Queries

Source: https://context7.com/brave/brave-search-cli/llms.txt

Use `bx spellcheck` to correct misspelled queries. The corrected query can be captured in a variable for subsequent use.

```bash
# Check and correct a misspelled query
bx spellcheck "kubernetse dployment" | jq '.results[0].query'
# → "kubernetes deployment"
```

```bash
# Use the corrected query in a pipeline
CORRECTED=$(bx spellcheck "asnc programing rust" | jq -r '.results[0].query')
bx context "$CORRECTED" --max-tokens 2048
```

--------------------------------

### Safe Search Control in Web Search

Source: https://context7.com/brave/brave-search-cli/llms.txt

Applies a strict safe search filter to web results, limiting the number of results to 10.

```bash
bx web "chemistry experiments" --safesearch strict --count 10
```

--------------------------------

### Filtering Result Types in Web Search

Source: https://context7.com/brave/brave-search-cli/llms.txt

Limits the search results to specific types, such as web pages and discussions.

```bash
bx web "rust error handling" --result-filter "web,discussions"
```

--------------------------------

### Non-Streaming AI Answer

Source: https://context7.com/brave/brave-search-cli/llms.txt

Obtains a single, non-streamed JSON response for an AI-grounded answer using the `--no-stream` flag. The output is piped to `jq .` for pretty-printing.

```bash
bx answers "compare SQLx and Diesel for Rust database access" --no-stream | jq .
```

--------------------------------

### Excluding Sites from News Search

Source: https://context7.com/brave/brave-search-cli/llms.txt

Performs a news search while excluding results from a specific domain, using past week freshness.

```bash
bx news "kubernetes CVE" --exclude-site medium.com --freshness pw
```

=== COMPLETE CONTENT === This response contains all available snippets from this library. No additional content exists. Do not make further requests.