Skip to main content
Glama

search_advanced

Search Perplexity across web, academic, and social sources with claim-level citations, then optionally verify each cited fact against its source page. Returns provenance, dates, and verifiable snippets.

Instructions

Search Perplexity.ai with specific source selection (web, academic, social), claim-level citation provenance ([1] markers, snippets, dates), optional live claim-source verification, local caching, cancellation-safe streaming, progress notifications, partial .txt checkpoints, typed error JSON, poolable browser profiles, and optional official Agent API backend.

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "query": "<query>",
      "sources": "<sources>"
    },
    "name": "search_advanced"
  }
}

Input Schema

NameRequiredDescriptionDefault
modelNoOptional Perplexity model name (e.g. 'Best', 'GPT-6.1 Sol', 'Grok 4.7', 'Claude Sonnet 5.5', 'Gemini 4 Hydrogen', or 'sonar' for API backend). Omit to use the currently selected model.
queryYesThe search query
backendNoExecution backend: 'auto' (uses official Perplexity Agent API if PERPLEXITY_API_KEY is configured and thread_url is omitted, otherwise browser Pro session), 'browser' (Playwright Pro UI), or 'api' (official HTTPS Agent API).
profileNoOptional browser profile name or directory path for isolated/concurrent execution (e.g. 'default', 'worker-2'). When omitted, automatically routes to an idle slot in PERPLEXITY_PROFILE_DIRS / default profile.
sourcesYesSources to search: 'web' (general web), 'academic' (scholarly articles), 'social' (Reddit & forums). Can combine multiple.
no_cacheNoIf true, bypasses the local response cache (default TTL 3600s, configurable via PERPLEXITY_CACHE_TTL_SECONDS) and forces a fresh search.
thread_urlNoOptional Perplexity thread URL returned by a previous search to ask a follow-up in the same conversation (browser backend).
output_fileNoOptional .txt file path to save the full answer to (regardless of length). Relative paths resolve inside the server's outputs directory. When omitted, answers at or above 15000 chars (configurable via PERPLEXITY_LONG_OUTPUT_CHARS) are automatically saved to a .txt file with an inline preview and path.
verify_claimsNoIf true (or PERPLEXITY_VERIFY_CLAIMS=1), fetches cited pages and checks key facts (numbers, dates, versions, names) for each claim, marking supported / weak / unverified.
timeout_secondsNoOptional timeout in seconds for waiting on the answer (defaults to server timeout, 900s / 15 minutes; max 1800s for very large code or multi-part responses).

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observedv1.4.0

TDQS

A3.7/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations, the description carries the full disclosure burden and does substantial work: it reveals local caching, live verification, cancellation-safe streaming, progress notifications, partial `.txt` checkpoints, typed error JSON, profile pooling, and an optional API backend. Authentication/session prerequisites and cache-freshness or default-backend behavior are left to the schema rather than highlighted as operational considerations.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The core action is front-loaded and a single roughly forty-word sentence is economical for a complex tool. However, the dense feature inventory mixes selection-critical capabilities with operational details, making it harder to scan than a short purpose statement followed by prioritized behavioral notes.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

The rich input schema covers invocation details, while the description usefully supplements it with citation provenance and runtime/failure behavior despite the absence of an output schema. It is sufficient for most calls, but explicit differentiation from `search`, authentication prerequisites, and a clearer success-result shape would make it more complete for this complexity.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, so the baseline is 3 and the schema already documents all ten arguments in greater detail. The description groups `sources`, `verify_claims`, `no_cache`, `profile`, and `backend` into a useful capability overview, but largely repeats their schema semantics rather than adding new defaults, syntax, or constraints.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

The opening “Search Perplexity.ai” provides a clear verb and resource, while source selection and claim-level citation provenance identify its advanced capabilities. It does not explicitly distinguish the tool from the sibling `search`, so the description alone does not establish which tool should handle an ordinary query.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines3/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Specific sources and optional claim verification imply use for controlled or higher-assurance searches, but this guidance is only implicit. There is no explicit when-to-use, when-not-to-use, or sibling-routing advice, particularly for choosing between `search_advanced` and `search`.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.