{"id":107224,"date":"2026-07-27T19:48:36","date_gmt":"2026-07-27T19:48:36","guid":{"rendered":"https:\/\/youzum.net\/perplexity-releases-pplx-a-single-binary-cli-that-puts-its-search-api-in-the-terminal-for-coding-agents\/"},"modified":"2026-07-27T19:48:36","modified_gmt":"2026-07-27T19:48:36","slug":"perplexity-releases-pplx-a-single-binary-cli-that-puts-its-search-api-in-the-terminal-for-coding-agents","status":"publish","type":"post","link":"https:\/\/youzum.net\/es\/perplexity-releases-pplx-a-single-binary-cli-that-puts-its-search-api-in-the-terminal-for-coding-agents\/","title":{"rendered":"Perplexity Releases pplx, a Single-Binary CLI That Puts Its Search API in the Terminal for Coding Agents"},"content":{"rendered":"

Perplexity<\/a> has released pplx<\/code>, an official command line client for its Search API. The tool returns grounded search results and extracted page text, all as JSON. According to its docs, it targets humans and coding agents equally. It is not a chat client. There is no conversational mode, no model selection and no synthesized answer.<\/p>\n

Two surfaces, one output contract<\/strong><\/h2>\n

The tool exposes exactly two working surfaces. pplx search web<\/code> runs a live web search. pplx content fetch<\/code> pulls a URL and returns cleaned page text.<\/p>\n

The contract around them is the interesting part. Per the official pplx-cli<\/code> Agent Skill<\/a>, success means exit code 0 and exactly one JSON object on stdout. Search returns {hits: [{url, title, domain, snippet, ...}], total, saved_to?}<\/code>.<\/p>\n

Every failure exits 1 with an empty stdout. One JSON error object goes to stderr, shaped {\"error\":{\"code\",\"message\",\"command\",\"hint\"?}}<\/code>. Documented codes include AUTHENTICATION<\/code>, UNKNOWN_ARGUMENT<\/code>, ARGUMENT_ERROR<\/code> and BAD_REQUEST<\/code>. The skill notes that list is not exhaustive, so callers should branch on error.code<\/code>.<\/p>\n

\n<\/div>\n

\n

Install path and platform support<\/strong><\/h2>\n<\/p>

Installation is a single shell command that pipes a script into sh<\/code>:<\/p>\n

\n
\n
\n
\n
<\/div>\n
<\/div>\n
<\/div>\n<\/div>\n
Copy Code<\/span>Copied<\/span>Use a different Browser<\/span><\/a><\/div>\n<\/div>\n
curl -fsSL https:\/\/github.com\/perplexityai\/perplexity-cli\/releases\/latest\/download\/install.sh | sh<\/code><\/pre>\n<\/div>\n<\/div>\n
Reading install.sh<\/code><\/a> shows what it actually does. It downloads manifest.json<\/code> from the latest release and extracts the tag and version. It then pins every remaining download to that tag, explicitly to avoid racing a concurrent publish.<\/p>\n
It fetches SHA256SUMS<\/code> and the platform binary, verifies the checksum, and installs to ~\/.local\/bin\/pplx<\/code>. No sudo is required. A receipt is written to ~\/.config\/pplx\/pplx-receipt.json<\/code>, but only after the installed binary runs successfully.<\/p>\n
Platform coverage is limited to three targets: macOS on Apple Silicon, Linux x86_64 and Linux arm64. Anything else exits with an error. There is no Windows build and no Intel macOS build.<\/p>\n
Context-window economics are a first-class design concern<\/strong><\/h2>\n
The most agent-specific feature is token budgeting. --output-dir<\/code> writes the full result set to a JSON file. --stdout-preview[=<CHARS>]<\/code> truncates long string fields in stdout, adding ...<truncated><\/code> markers.<\/p>\n
The skill is blunt about the trap: --stdout-preview<\/code> is a no-op without a save directory. It truncates only when the result is also saved via --output-dir<\/code> or $PPLX_OUTPUT_DIR<\/code>. Used alone it returns full-size output, and hits can be multiple KB each.<\/p>\n
Saved search results land at {dir}\/web\/{rand}.json<\/code> and fetches at {dir}\/fetch\/{rand}.json<\/code>. Files are written only after a successful request. PPLX_OUTPUT_DIR<\/code> sets a workspace default so the flag need not be repeated.<\/p>\n
For content fetch, the skill adds a correctness check rather than a cost one. Verify error<\/code> and is_paywall<\/code> in the output before trusting content<\/code>. --html<\/code> adds a raw_html<\/code> field fetched live via crawler, and --no-cache<\/code> forces a live fetch. <\/p>\n
\n<\/div>\n
\n
Key Takeaways<\/strong><\/h2>\n