Skip to content
verify mcp Beta VerifyMCP is currently in beta. If you notice any issues, get in touch and we’ll put it right.

com.mcparmory/perplexity

PYPI · MCPARMORY-PERPLEXITY · 2 COMPONENTS · SCANNED SEP 20

Search the web, generate AI responses, and create embeddings with real-time information

0 this week 73 Trust /100
Trust breakdown (7 categories)

How this component scores in each security and reliability category. Every signal is checked automatically from public evidence about the published package, including repeated runs of it in an isolated sandbox, and we only credit what we can confirm. How we score → Why this is hard to score →

Supply Chain Security87
  • No malware found by supply-chain analysis.Pass
  • No known CVEs affecting this package version or its production dependencies.Pass
  • Install-script risk not yet assessed.Unverified
  • 2 of 41 dependencies flagged as unhealthy. View diagnostics → Partial
Provenance & Transparency32
Schema Quality & AI Usability60
  • AI-judged instruction clarity (excellent).Pass
  • Context-footprint check failed: tool/resource definitions use about 2755 tokens (~393/item across 7 items; 7 tools + 0 resources), over budget; trim descriptions and params. See how to fix → Fail
  • Usage-examples check failed: none of the tools include examples. See how to fix → Fail
Stability & Change Management87
  • Stability observed for 26 of 30 days with no destabilising changes; credit accrues until the full window elapses.Partial
Tool Coverage100
  • 100% of tools have a non-trivial description (not blank, and not just the tool's name).Pass
  • 100% of tool parameters carry a description.Pass
Tool Safety100
  • No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.Pass
  • We read all 7 captured tool definition(s), and no name or description among them implies an irreversible operation.Pass
  • An AI judge read all 7 captured unit(s) of tool text and found none that tries to manipulate the model reading it.Pass
Capabilities100
  • Implements a current MCP spec version (2026-07-28).Pass
Install

How do I install the com.mcparmory/perplexity MCP server?

com.mcparmory/perplexity runs locally as a PyPI package, launched with uvx mcparmory-perplexity. Ready-made configuration for Claude, Cursor, VS Code, Codex and 5 more is on this page, copied from each client's own documentation.

pypi · mcparmory-perplexity

# add to Claude Code
claude mcp add com-mcparmory-perplexity -- uvx mcparmory-perplexity
// .cursor/mcp.json
{
  "mcpServers": {
    "com-mcparmory-perplexity": {
      "command": "uvx",
      "args": [
        "mcparmory-perplexity"
      ]
    }
  }
}
// .vscode/mcp.json
{
  "servers": {
    "com-mcparmory-perplexity": {
      "command": "uvx",
      "args": [
        "mcparmory-perplexity"
      ]
    }
  }
}
# add to Codex CLI
codex mcp add com-mcparmory-perplexity -- uvx mcparmory-perplexity
// opencode.json
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "com-mcparmory-perplexity": {
      "type": "local",
      "command": [
        "uvx",
        "mcparmory-perplexity"
      ],
      "enabled": true
    }
  }
}
# add to OpenClaw
openclaw mcp add com-mcparmory-perplexity --command uvx --arg mcparmory-perplexity
# ~/.hermes/config.yaml
mcp_servers:
  com-mcparmory-perplexity:
    command: "uvx"
    args: ["mcparmory-perplexity"]
// ~/.netclaw/config/netclaw.json
{
  "McpServers": {
    "com-mcparmory-perplexity": {
      "Transport": "stdio",
      "Command": "uvx",
      "Arguments": [
        "mcparmory-perplexity"
      ]
    }
  }
}
# add to Vellum
assistant mcp add com-mcparmory-perplexity -t stdio -c uvx -a mcparmory-perplexity
// mcp.json
{
  "mcpServers": {
    "com-mcparmory-perplexity": {
      "command": "uvx",
      "args": [
        "mcparmory-perplexity"
      ]
    }
  }
}
Changelog

Every change we have recorded for this component, newest first. Security-relevant changes are always shown. ▲ marks a change for the better, ▼ a change for the worse; unmarked changes are neutral.

  • 18 Sept 26 −2
    • Stability: pass → 0.80 functional
  • 17 Sept 26 0
    • Stability: 0.97 → pass security
  • 16 Sept 26 +16
    • Malware scan: unverified → pass security
  • 14 Sept 26 −14
    • Malware scan: pass → unverified security
  • 11 Sept 26 −2
    • Stability: pass → 0.80 functional
  • 10 Sept 26 0
    • Stability: 0.97 → pass security
  • 9 Sept 26 +1

    No change was recorded against any check on this day. Stability & Change Management went from 93 to 97. That category is still filling its 30-day observation window: 28 days of observed history at the previous scan, 29 at this one. The score rises as the window fills, whether or not the server changes.

  • 7 Sept 26 +1

    No change was recorded against any check on this day. Stability & Change Management went from 87 to 90. That category is still filling its 30-day observation window: 26 days of observed history at the previous scan, 27 at this one. The score rises as the window fills, whether or not the server changes.

Diagnostics

Diagnostic detail from the automated scan of this channel: what the scanner observed at each step, so you can see exactly where a check passed or failed. It is informational only and never changes the trust score.

Captured 20 Sept 2026 · Analysed pypi/mcparmory-perplexity@1.0.2

Provenance No attestation

The registry publishes no build provenance for this version, so there is nothing to verify.

Result No attestation
Ecosystem pypi

Background: How many MCP packages publish verified provenance →

Dependencies 41 packages
Packages resolved 41
Stale 1
No linked repository 1
Tree resolution Complete

Background: SBOMs and build attestations, explained →

MCP tools · 7 exposed · ~2,755 tokens

The tools this component advertises to a client, with an estimated token cost for each. Expand a tool to see its parameters and schema. The per-tool counts are indicative and are not scored directly; the schema's total context footprint is one signal in Schema Quality & AI Usability. A tool's description is untrusted text the model reads on every call, which is what makes this list a security surface and not just an inventory: how tool poisoning works →

Tool Tokens
create_async_chat_completion ~1,087

Submit an asynchronous chat completion request with optional web search, structured output, and streaming capabilities. Supports multiple model variants and fine-grained control over search behavior, response format, and reasoning depth.

NameTypeReqDescription
citystringCity name to provide geographic context for location-aware search and response localization.
countrystringISO 3166-1 alpha-2 country code to provide geographic context for search results and localized responses.
disable_searchbooleanWhen enabled, completely disables web search functionality and the model responds exclusively based on its training data without accessing external sources.
enable_search_classifierbooleanWhen enabled, uses a classifier to intelligently determine whether web search is necessary for the query or if the model's training data is sufficient.
image_domain_filterarrayRestrict image results to specific domains by providing a list of domain names. Images will only be sourced from these domains.
image_format_filterarrayFilter image results by file format. Provide a list of formats (e.g., png, jpg, gif, webp) to include only images in those formats.
image_results_enhanced_relevancebooleanWhen enabled, applies enhanced relevance filtering to image results to prioritize higher-quality and more contextually relevant images.
language_preferencestringSpecify the preferred response language using an ISO 639-1 language code (e.g., en, es, fr, de). The model will attempt to respond in this language when possible.
latitudenumberLatitude coordinate for location-based search context. Use with longitude, city, region, or country to refine geographic relevance.
longitudenumberLongitude coordinate for location-based search context. Use with latitude, city, region, or country to refine geographic relevance.
max_tokensintegerMaximum number of tokens to generate in the completion response. Must be greater than 0 and cannot exceed 128,000 tokens.
messagesarrayyesArray of message objects forming the conversation history. Each message should include role (user, assistant, system) and content fields.
modelstringyesThe model variant to use for processing the request. Choose from sonar (base), sonar-pro (enhanced), sonar-deep-research (comprehensive), or sonar-reasoning-pro (advanced reasoning).
reasoning_effortstringControls the computational effort allocated to reasoning. Choose 'minimal' for quick responses, 'low' or 'medium' for balanced reasoning, or 'high' for deep analytical reasoning.
regionstringState or region name to provide geographic context for location-aware search and response localization.
response_formatSpecifies the output format structure. Set type to 'json_schema' to enforce structured JSON output; omit for default text responses.
return_imagesbooleanWhen enabled, includes relevant images in the response alongside text results.
return_related_questionsbooleanWhen enabled, generates suggested follow-up questions based on the search results to facilitate deeper exploration of the topic.
search_context_sizestringControls the breadth of search context included in the response. Choose 'low' for minimal context, 'medium' for balanced coverage, or 'high' for comprehensive search results. Defaults to 'low'.
search_domain_filterarrayRestrict search results to specific domains by providing a list of domain names (e.g., github.com, wikipedia.org). Results will only include pages from these domains.
search_language_filterarrayFilter search results by language using ISO 639-1 language codes (e.g., en for English, fr for French, de for German). Provide as a list of codes.
search_modestringSpecifies the source domain for search results. Choose 'web' for general internet search, 'academic' for scholarly articles, or 'sec' for SEC filings and financial documents.
search_recency_filterstringFilter search results by publication recency. Choose from 'hour' (last hour), 'day' (last 24 hours), 'week' (last 7 days), 'month' (last 30 days), or 'year' (last 365 days).
search_typestringDetermines search quality versus speed tradeoff. Use 'fast' for quick results, 'pro' for higher quality results, or 'auto' to let the model decide based on the query.
stopOne or more stop sequences that will terminate generation when encountered. Provide as a list of strings.
streambooleanEnable server-sent event (SSE) streaming to receive response chunks in real-time as they are generated.
stream_modestringControls the format of streaming events when streaming is enabled. Use 'full' to suppress reasoning events and include metadata inline; use 'concise' to emit reasoning events separately. Defaults to…
temperaturenumberControls response randomness on a scale from 0 to 2. Lower values (closer to 0) produce more deterministic outputs; higher values increase creativity and variability.
top_pnumberNucleus sampling parameter between 0 and 1 that controls output diversity. Lower values focus on the most likely tokens; higher values allow more diverse token selection.

No output schema declared.

No examples provided.

create_chat_completion ~961

Generate a chat completion response using a specified Sonar model, with optional web search, streaming, and structured output capabilities. Supports conversation history, location context, and advanced search filtering.

NameTypeReqDescription
citystringCity name to provide geographic context for search results and localized responses.
countrystringISO 3166-1 alpha-2 country code (e.g., US, GB, FR) to localize search results and responses.
disable_searchbooleanDisable all web search capabilities and respond based exclusively on the model's training data.
enable_search_classifierbooleanEnable automatic classification to determine whether web search is necessary for the query, optimizing performance for queries answerable from training data alone.
image_domain_filterarrayArray of domain names to limit image results to specific sources. Order is not significant.
image_format_filterarrayArray of image file formats (e.g., png, jpg, gif) to filter image results. Order is not significant.
image_results_enhanced_relevancebooleanWhen enabled, applies enhanced relevance filtering to image results to prioritize higher-quality matches.
language_preferencestringISO 639-1 language code (e.g., en, fr, de) to specify the preferred language for the response.
latitudenumberLatitude coordinate for location-aware search results, typically used with longitude, city, region, or country.
longitudenumberLongitude coordinate for location-aware search results, typically used with latitude, city, region, or country.
max_tokensintegerMaximum number of tokens to generate in the completion, between 1 and 128,000.
messagesarrayyesArray of message objects representing the conversation history, with each message containing role and content. Order matters—earlier messages provide context for later responses.
modelstringyesThe model variant to use for generating the response. Choose from sonar (base), sonar-pro (enhanced), sonar-deep-research (comprehensive), or sonar-reasoning-pro (advanced reasoning).
reasoning_effortstringControls reasoning depth for models that support it: minimal (fastest), low (basic), medium (balanced), or high (most thorough). Only applicable to reasoning-capable models.
regionstringState or region name to refine geographic context for search results.
response_formatSpecifies output format constraints. Set type to json_schema for structured JSON output, or omit for default text responses.
return_imagesbooleanInclude image results in the response alongside text results.
return_related_questionsbooleanGenerate suggested follow-up questions based on the search results to facilitate deeper exploration of the topic.
search_context_sizestringControls the breadth of search context included in the response: low (minimal), medium (balanced), or high (comprehensive). Defaults to low.
search_domain_filterarrayArray of domain names (e.g., github.com, wikipedia.org) to limit search results to specific sources. Order is not significant.
search_language_filterarrayArray of ISO 639-1 language codes (e.g., en, fr, de) to filter results by language. Order is not significant.
search_modestringRestrict search results to a specific source: web (general internet), academic (scholarly articles), or sec (SEC filings and financial documents).
search_recency_filterstringFilter search results by publication recency: hour (last hour), day (last 24 hours), week (last 7 days), month (last 30 days), or year (last 365 days).
search_typestringSearch quality strategy: fast prioritizes speed, pro prioritizes result quality, auto lets the model decide based on the query.
stopOne or more strings that will cause generation to stop immediately when produced. Useful for controlling output length or format.
streambooleanEnable server-sent event streaming to receive the response incrementally rather than waiting for the complete result.
stream_modestringControls streaming event format: full suppresses reasoning events and includes metadata inline for cleaner output; concise emits reasoning events separately for detailed insight. Defaults to full.
temperaturenumberControls response randomness on a scale from 0 (deterministic) to 2 (highly random). Lower values produce more consistent outputs; higher values increase creativity.
top_pnumberNucleus sampling parameter between 0 and 1 that controls output diversity by considering only the most likely tokens that sum to this probability.

No output schema declared.

No examples provided.

create_contextualized_embeddings ~284

Generate contextualized embeddings for document chunks where chunks from the same document share context awareness, improving retrieval quality for document-based applications.

NameTypeReqDescription
dimensionsintegerOptional number of dimensions for the output embeddings using Matryoshka scaling. Supported range is 128 to 1024 dimensions for the 0.6B model and 128 to 2560 dimensions for the 4B model. Defaults to…
encoding_formatstringOptional output encoding format for embeddings. Use base64_int8 for base64-encoded signed 8-bit integers, or base64_binary for base64-encoded packed binary format with 1 bit per dimension. Defaults t…
inputarrayyesNested array where each inner array contains text chunks from a single document. Chunks within the same document are encoded together to maintain document-level context awareness. Supports up to 512…
modelstringyesThe contextualized embedding model to use. Choose between the 0.6B parameter model or the 4B parameter model based on your accuracy and performance requirements.

No output schema declared.

No examples provided.

create_embeddings ~266

Generate vector embeddings for one or more texts to enable semantic search, clustering, and machine learning applications. Supports flexible output dimensions and encoding formats.

NameTypeReqDescription
dimensionsintegerOptional number of dimensions for the output embeddings using Matryoshka scaling. For pplx-embed-v1-0.6b, specify between 128 and 1024 dimensions; for pplx-embed-v1-4b, between 128 and 2560 dimension…
encoding_formatstringOptional output encoding format for embeddings. Use base64_int8 for base64-encoded signed 8-bit integers, or base64_binary for base64-encoded packed binary (1 bit per dimension). Defaults to base64_i…
inputyesText or texts to embed as a string or array of strings. Submit up to 512 texts per request, with each text limited to 32K tokens and a combined total of 120,000 tokens across all inputs. Empty string…
modelstringyesThe embedding model to use. Choose between pplx-embed-v1-0.6b (1024 dimensions) or pplx-embed-v1-4b (2560 dimensions).

No output schema declared.

No examples provided.

get_async_chat_completion_response ~50

Retrieve the completion response for a previously submitted asynchronous chat request using its request identifier.

NameTypeReqDescription
api_requeststringyesThe unique identifier of the asynchronous chat completion request whose response you want to retrieve.

No output schema declared.

No examples provided.

list_async_chat_completions ~42

Retrieve a list of all asynchronous chat completion requests submitted by the authenticated user. This endpoint provides visibility into the status and history of async chat completion jobs.

Input schema present but exposes no named parameters.

No output schema declared.

No examples provided.

search_web ~65

Search the web and retrieve relevant content from web pages matching your query. Returns a curated list of results with page contents.

NameTypeReqDescription
bodyobjectyesSearch query parameters including the search terms, optional filters, and result preferences. Specify your search query and any desired constraints for filtering results.

No output schema declared.

No examples provided.

Common questions

What is the com.mcparmory/perplexity MCP server?

com.mcparmory/perplexity is an MCP server listed in the public MCP registry as com.mcparmory/perplexity. Search the web, generate AI responses, and create embeddings with real-time information. This page covers its PyPI package (mcparmory-perplexity).

Is the com.mcparmory/perplexity MCP server safe to use?

com.mcparmory/perplexity scores 73 out of 100 on VerifyMCP. We found no known CVEs affecting it as of 20 September 2026. That is a record of what we were able to check automatically, not an endorsement. The category breakdown on this page shows every signal behind the number, including the ones we could not confirm.

What tools does the com.mcparmory/perplexity MCP server expose?

com.mcparmory/perplexity exposes 7 tools: create_chat_completion, search_web, create_embeddings, create_contextualized_embeddings, get_async_chat_completion_response, and 2 more. Their descriptions and schemas cost roughly 2,755 tokens of context every time the server is loaded.

Is the com.mcparmory/perplexity MCP server still maintained?

com.mcparmory/perplexity is still listed as active in the MCP registry. We last reached this channel on 20 September 2026. Those dates come from our own scans of the registry and the channel itself, not from anything the publisher announced.