> ## Documentation Index
> Fetch the complete documentation index at: https://composio-preview.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Perplexity AI

> Perplexity AI provides conversational AI models for generating human-like text responses

![Perplexity AI logo](https://mintcdn.com/composio-preview/gqjBKmmJCINafaiK/images/toolkits/perplexityai.svg?fit=max\&auto=format\&n=gqjBKmmJCINafaiK\&q=85\&s=5d2c7f5b5cbba7a3ad4d86f50a318da0)

* **Category:** artificial intelligence
* **Auth:** API\_KEY
* **Composio-managed OAuth available?** N/A
* **Tools:** 28
* **Triggers:** 0
* **Slug:** `PERPLEXITYAI`
* **Version:** 20261101\_00

## Tools

> [accordion] Cancel Agent Response
>
> **Slug:** `PERPLEXITYAI_CANCEL_AGENT_RESPONSE`
>
> Request cancellation of an existing durable agent response. A successful acknowledgement means cancellation was accepted, not that the run has stopped. Poll GET /v1/agent/\{id} for its terminal status. An already terminal run returns HTTP 400; unknown or cross-account IDs return HTTP 404.
>
> #### Input Parameters
>
> **id** `string` *required*
>
> Response ID to cancel (resp\_\<...>).
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] Create Async Chat Completion
>
> **Slug:** `PERPLEXITYAI_CREATE_ASYNC_CHAT_COMPLETION`
>
> Create Async Chat Completion (POST /v1/async/sonar). Submits an asynchronous chat completion request for long-running tasks. Returns immediately with a request ID that can be polled using the Get Async Chat Completion action. Only the 'sonar-deep-research' model is supported for async processing. Async jobs have a 7-day TTL. Deep research generates very long responses (10K-100K+ words) with exhaustive multi-source analysis. Use the idempotency\_key to prevent duplicate submissions. Poll with Get Async Chat Completion using the returned request ID to retrieve results when status is COMPLETED.
>
> #### Input Parameters
>
> **request** `object` *required*
>
> The chat completion request to run asynchronously.
>
> **idempotency\_key** `string`
>
> Optional key to prevent duplicate requests. Deduplication requires the exact same request body.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] Create Chat Completion
>
> **Slug:** `PERPLEXITYAI_CREATE_CHAT_COMPLETION`
>
> Generate a chat-compatible response from conversation messages, with optional search controls. Streaming responses are collected into a chat-completion result; recency and date filters are mutually exclusive. DEPRECATED: Sonar support has ended; prefer Execute Agent for new integrations.
>
> #### Input Parameters
>
> **stop** `string`
>
> Stop sequence(s) to end generation. Note: the API may not support custom stop words for all models.
>
> **model** `string ("sonar" | "sonar-pro" | "sonar-deep-research" | "sonar-reasoning-pro")`
>
> The Sonar model to use. 'sonar' for fast/cheap queries, 'sonar-pro' for complex queries, 'sonar-reasoning-pro' for chain-of-thought reasoning, 'sonar-deep-research' for exhaustive research (generates very long 10K+ word responses; prefer the async endpoint for this model).
>
> **top\_p** `number`
>
> Nucleus sampling threshold controlling cumulative probability of tokens.
>
> **stream** `boolean`
>
> Whether to stream the response using server-sent events.
>
> **messages** `array` *required*
>
> A list of messages comprising the conversation. Each message has a role (system, user, assistant, tool) and content. Messages with 'tool' role must alternate with 'assistant' messages. Valid pattern: system -> user -> assistant -> tool -> user.
>
> **max\_tokens** `integer`
>
> Maximum number of tokens to generate. Max 128000.
>
> **search\_mode** `string`
>
> Search mode. Options: 'web', 'academic', 'sec'.
>
> **stream\_mode** `string`
>
> Stream mode: 'full' (default) or 'concise'.
>
> **temperature** `number`
>
> Controls randomness. 0 is deterministic, up to 2 is more random.
>
> **return\_images** `boolean`
>
> Whether to include images in the response.
>
> **disable\_search** `boolean`
>
> Whether to disable web search entirely.
>
> **response\_format** `object`
>
> Response format configuration.
>
> **reasoning\_effort** `string`
>
> Reasoning effort level: 'minimal', 'low', 'medium', 'high'.
>
> **web\_search\_options** `object`
>
> Options for web search behavior.
>
> **image\_domain\_filter** `array`
>
> Filter images by source domain.
>
> **image\_format\_filter** `array`
>
> Filter images by format (e.g. 'png', 'jpg').
>
> **language\_preference** `string`
>
> Preferred response language as ISO 639-1 code (e.g. 'en', 'fr').
>
> **search\_domain\_filter** `array`
>
> List of domains to restrict search results to.
>
> **search\_recency\_filter** `string`
>
> Filter by recency: 'hour', 'day', 'week', 'month', 'year'. Mutually exclusive with date filters (search\_after\_date\_filter, search\_before\_date\_filter, last\_updated\_after\_filter, last\_updated\_before\_filter).
>
> **search\_language\_filter** `array`
>
> ISO 639-1 language codes to filter search results.
>
> **enable\_search\_classifier** `boolean`
>
> Whether to enable the search classifier.
>
> **return\_related\_questions** `boolean`
>
> Whether to return related questions.
>
> **search\_after\_date\_filter** `string`
>
> Only include results published after this date (MM/DD/YYYY). Cannot be used with search\_recency\_filter.
>
> **last\_updated\_after\_filter** `string`
>
> Only include results last updated after this date (MM/DD/YYYY).
>
> **search\_before\_date\_filter** `string`
>
> Only include results published before this date (MM/DD/YYYY). Cannot be used with search\_recency\_filter.
>
> **last\_updated\_before\_filter** `string`
>
> Only include results last updated before this date (MM/DD/YYYY).
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] Create Contextualized Embeddings
>
> **Slug:** `PERPLEXITYAI_CREATE_CONTEXTUALIZED_EMBEDDINGS`
>
> Create Contextualized Embeddings (POST /v1/contextualizedembeddings). Generates document-aware embeddings where chunks from the same document share context. Unlike standard embeddings, these recognize sequential relationships within documents, improving retrieval quality. Models: pplx-embed-context-v1-0.6b (1024 dims) and pplx-embed-context-v1-4b (2560 dims). Both support Matryoshka dimension reduction and INT8/binary quantization.
>
> #### Input Parameters
>
> **input** `array` *required*
>
> Nested array of texts: outer array = documents, inner array = ordered chunks from that document. Max 512 documents, max 16000 total chunks, max 32K tokens per document, max 120K combined tokens. Empty strings not allowed.
>
> **model** `string ("pplx-embed-context-v1-0.6b" | "pplx-embed-context-v1-4b")` *required*
>
> The contextualized embedding model. pplx-embed-context-v1-0.6b (1024 dims, \$0.008/1M tokens) or pplx-embed-context-v1-4b (2560 dims, \$0.05/1M tokens).
>
> **dimensions** `integer`
>
> Custom output dimensions via Matryoshka. Range: 128-1024 for 0.6b, 128-2560 for 4b. Defaults to full dimensions.
>
> **encoding\_format** `string ("base64_int8" | "base64_binary")`
>
> Encoding format for embeddings output.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] CreateEmbeddings
>
> **Slug:** `PERPLEXITYAI_CREATE_EMBEDDINGS`
>
> Generate vector embeddings for independent texts (queries, sentences, documents). This action takes one or more input texts and generates vector embeddings using Perplexity AI's embedding models. Embeddings are useful for semantic search, similarity matching, and machine learning downstream tasks. Supported models: - pplx-embed-v1-0.6b: Smaller, faster model (1024 dimensions) - pplx-embed-v1-4b: Larger, more accurate model (2560 dimensions) The output embeddings are base64-encoded for efficient transmission. Use the dimensions parameter to reduce embedding size for faster processing when full precision is not required (Matryoshka representation).
>
> #### Input Parameters
>
> **input** `string` *required*
>
> Input text to embed, encoded as a string or array of strings. Maximum 512 texts per request. Each input must not exceed 32K tokens. All inputs in a single request must not exceed 120,000 tokens combined. Empty strings are not allowed.
>
> **model** `string ("pplx-embed-v1-0.6b" | "pplx-embed-v1-4b")` *required*
>
> The embedding model to use. pplx-embed-v1-0.6b is smaller and faster, while pplx-embed-v1-4b is larger and more accurate.
>
> **dimensions** `integer`
>
> Number of dimensions for output embeddings (Matryoshka). Range: 128-1024 for pplx-embed-v1-0.6b, 128-2560 for pplx-embed-v1-4b. Defaults to full dimensions (1024 or 2560).
>
> **encoding\_format** `string ("base64_int8" | "base64_binary")`
>
> Encoding format for embeddings output.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] Create Orchestrator Chat Completion
>
> **Slug:** `PERPLEXITYAI_CREATE_ORCHESTRATOR_CHAT_COMPLETION`
>
> Create a non-streaming OpenAI-compatible chat completion using a cataloged Orchestrator model. Supports multimodal conversations, structured output, reasoning, caching and tool calls where the selected model supports them. Returns the completion and token accounting; stream=True is unsupported.
>
> #### Input Parameters
>
> **stop** `string`
>
> One stop sequence or an array of 1 to 4 stop sequences.
>
> **user** `string`
>
> Deprecated SDK-compatibility user identifier; not forwarded to the model.
>
> **model** `string` *required*
>
> Catalog model id, including creator prefix, e.g. perplexity/kimi-k3.
>
> **tools** `array`
>
> Function or custom tool definitions. Every tool requires a description. Function tools use type=function and function with name, parameters (JSON Schema), nullable strict; custom tools use type=custom and custom with name and optional text or grammar format (definition and lark/regex syntax).
>
> **top\_p** `number`
>
> Nucleus sampling threshold, 0 to 1; adjust this or temperature, not both.
>
> **stream** `boolean`
>
> Only false or null is supported. True (server-sent events) is rejected by this non-streaming action; omitted values are not sent.
>
> **messages** `array` *required*
>
> Conversation messages with role (developer, system, user, assistant, tool, or legacy function), content and optional name. Content may be text or multimodal text/image\_url/input\_audio/file parts; assistant messages may include nullable content/refusal/audio, reasoning\_content, tool\_calls or legacy function\_call; tool messages require tool\_call\_id.
>
> **metadata** `object`
>
> String metadata map accepted for SDK compatibility; not forwarded to the model.
>
> **max\_tokens** `integer`
>
> Deprecated legacy alias of max\_completion\_tokens.
>
> **temperature** `number`
>
> Sampling temperature, 0 to 2; adjust this or top\_p, not both.
>
> **tool\_choice** `string`
>
> Tool selection mode, or named function/custom choice, or type=allowed\_tools with allowed\_tools containing mode auto/required and tools array.
>
> **service\_tier** `string ("auto" | "default" | "flex" | "scale" | "priority")`
>
> Processing tier. Flex halves rates on supported models; auto and default are equivalent.
>
> **stream\_options** `object`
>
> Streaming settings with optional include\_usage boolean. Provider only allows this when stream=true; streaming is unsupported by this action.
>
> **response\_format** `object`
>
> Output format with type text, json\_object, or json\_schema. For json\_schema include json\_schema with required name, optional schema (extensible JSON Schema) and nullable strict.
>
> **prompt\_cache\_key** `string`
>
> Stable key grouping related requests to improve prompt-cache hits.
>
> **reasoning\_effort** `string ("none" | "minimal" | "low" | "medium" | "high" | "xhigh" | "max")`
>
> Reasoning effort on supported models; omission uses the model/provider default.
>
> **safety\_identifier** `string`
>
> SDK-compatibility safety identifier; not forwarded to the model.
>
> **parallel\_tool\_calls** `boolean`
>
> Whether multiple tool calls may be requested in one turn.
>
> **prompt\_cache\_options** `object`
>
> Prompt-cache settings: ttl and mode (implicit or explicit). Saved schema lists ttl=30m while operation prose lists 5m, 1h, 24h; forwarded unchanged for provider validation.
>
> **max\_completion\_tokens** `integer`
>
> Upper bound on generated tokens, including reasoning tokens.
>
> **prompt\_cache\_retention** `string ("in_memory" | "24h")`
>
> Deprecated cache retention; prefer prompt\_cache\_options.ttl.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] Create Orchestrator Message
>
> **Slug:** `PERPLEXITYAI_CREATE_ORCHESTRATOR_MESSAGE`
>
> Create a non-streaming Anthropic-compatible message using an Orchestrator catalog model. Supply conversation history and max\_tokens; optionally configure tools, sampling, thinking, caching or structured output where the model supports them. Returns content blocks, stop information and usage. stream=True is unsupported.
>
> #### Input Parameters
>
> **model** `string` *required*
>
> Public catalog model slug, e.g. perplexity/kimi-k3; slashes are supported.
>
> **speed** `string ("standard" | "fast")`
>
> standard is inert; fast is rejected by the provider because fast-mode routing is unavailable.
>
> **tools** `array`
>
> Custom tools requiring name and input\_schema (raw JSON Schema). Optional type=custom, description, cache\_control, strict, defer\_loading; input\_examples, eager\_input\_streaming and allowed\_callers are accepted but ignored. Server tool types are rejected. Provider schemas are forwarded unchanged.
>
> **top\_k** `integer`
>
> Sample only from the top K options for each subsequent token.
>
> **top\_p** `number`
>
> Nucleus sampling probability from 0 to 1.
>
> **stream** `boolean`
>
> Omit or set false for JSON output. True requests server-sent events and is unsupported by this action.
>
> **system** `string`
>
> System prompt as text or text blocks with type=text, text and optional cache\_control.
>
> **messages** `array` *required*
>
> Conversation messages, each requiring role (user, assistant, system) and content (text or content-block array). Blocks include text, image (base64 or URL source), document (base64 PDF, text/plain, or URL source), search\_result, tool\_use (id, name, input), tool\_result (tool\_use\_id, optional content/is\_error), thinking (thinking/signature replay), and redacted\_thinking (rejected by provider). Document title/context may be null; non-base64 documents, title/context and search\_result require anthropic/ models. Citation configs are rejected. Provider objects are forwarded unchanged.
>
> **metadata** `object`
>
> Request metadata containing optional nullable user\_id string; accepted but not forwarded to the model.
>
> **thinking** `object`
>
> Thinking config: type enabled with required budget\_tokens (mapped to reasoning effort), adaptive (automatic thinking), or disabled. Enabled/adaptive may include display summarized or omitted. Support depends on provider/model; provider validates the object.
>
> **container** `string`
>
> Raw JSON container configuration. Rejected by provider validation: code-execution containers are unsupported.
>
> **fallbacks** `string`
>
> Raw JSON fallback configuration. Rejected by provider validation because fallback models would bill as the requested slug.
>
> **max\_tokens** `integer` *required*
>
> Maximum generated tokens (1 to 2147483647). Required by the documented Messages endpoint for all models, including Anthropic routing.
>
> **mcp\_servers** `string`
>
> Raw JSON MCP server configuration. Rejected by provider validation: no server-side MCP execution.
>
> **temperature** `number`
>
> Sampling temperature from 0 to 1.
>
> **tool\_choice** `object`
>
> Tool selection object: type auto, any, tool (requires name), or none. Auto/any/tool may include disable\_parallel\_tool\_use.
>
> **service\_tier** `string ("auto" | "standard_only")`
>
> Processing tier. Saved schema marks this accepted and ignored because the upstream tier is OpenAI-only; reference prose instead says rejected. Forwarded for provider validation.
>
> **cache\_control** `object`
>
> Cache breakpoint object requiring type=ephemeral and optional ttl 5m or 1h. The provider rejects 1h until distinctly billed; support varies by model.
>
> **inference\_geo** `string`
>
> Geographic inference pinning. Rejected by provider validation because residency cannot be honored.
>
> **output\_config** `object`
>
> Output configuration: format with type=json\_schema and schema (raw JSON Schema); effort low, medium, high, xhigh, or max; task\_budget is accepted but ignored.
>
> **stop\_sequences** `array`
>
> Sequences that stop generation.
>
> **context\_management** `object`
>
> Extensible context-edit strategy object; accepted and ignored, edits are not applied.
>
> **fallback\_credit\_token** `string`
>
> Fallback credit token; rejected by provider validation together with fallbacks.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] CreateOrchestratorResponse
>
> **Slug:** `PERPLEXITYAI_CREATE_ORCHESTRATOR_RESPONSE`
>
> Generate a non-streaming model response through the Perplexity Orchestrator using a creator/model-name catalog ID, input text or structured items, optional tools, and structured output settings. Returns the full Responses-compatible resource; stream=True is rejected.
>
> #### Input Parameters
>
> **text** `object`
>
> Text configuration: verbosity low/medium/high and format with type text or json\_schema. json\_schema requires name and schema, accepts description and nullable strict=true. JSON Schema is forwarded verbatim.
>
> **input** `string` *required*
>
> User text (maximum 10485760 characters), or input items: messages with role/content and optional phase/status/id; reasoning with summary/encrypted\_content; function/custom tool calls or outputs with call\_id; or additional\_tools. Content parts and provider-specific properties are forwarded verbatim.
>
> **model** `string` *required*
>
> Orchestrator catalog model ID in creator/model-name form; slashes are preserved.
>
> **tools** `array`
>
> Tools available to the model: function (name, description, parameters JSON Schema, strict), custom (name, description, format text/grammar), or namespace (name, description, tools). Include the type discriminator; nested schemas are forwarded verbatim.
>
> **top\_p** `number`
>
> Nucleus sampling cumulative probability between 0 and 1.
>
> **stream** `boolean`
>
> Must be false. This action returns a complete JSON response and does not support server-sent event streaming.
>
> **include** `array`
>
> Additional response data to include; reasoning.encrypted\_content enables replay of encrypted reasoning.
>
> **metadata** `object`
>
> Up to 16 metadata pairs; keys at most 64 characters and string values at most 512 characters.
>
> **reasoning** `object`
>
> Reasoning model configuration: effort none/low/medium/high/xhigh or null, and summary concise/detailed/auto. Model support varies.
>
> **truncation** `string`
>
> Only disabled is supported; context overflow produces a provider HTTP 400.
>
> **temperature** `number`
>
> Sampling temperature between 0 and 2.
>
> **tool\_choice** `string`
>
> Tool selection policy, or a specific function object with type='function' and name.
>
> **instructions** `string`
>
> Additional instructions guiding this request.
>
> **service\_tier** `string ("auto" | "default" | "flex" | "priority")`
>
> Requested service tier; availability depends on model and account.
>
> **prompt\_cache\_key** `string`
>
> Prompt cache read/write key, up to 64 characters.
>
> **max\_output\_tokens** `integer`
>
> Maximum generated response tokens; minimum 16.
>
> **safety\_identifier** `string`
>
> Stable safety monitoring and abuse detection identifier, up to 64 characters.
>
> **parallel\_tool\_calls** `boolean`
>
> Whether the model may make multiple tool calls in parallel.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] Create Skill
>
> **Slug:** `PERPLEXITYAI_CREATE_SKILL`
>
> Create a Project-owned skill from a ZIP bundle containing SKILL.md. The bundle supplies the skill name and description. Returns the skill ID and revision metadata for subsequent use, updates, or deletion. The provider may reject invalid bundles, oversized uploads, or a Project's skill limit.
>
> #### Input Parameters
>
> **file** `object` *required*
>
> ZIP skill bundle containing SKILL.md. The skill name and description are read from SKILL.md, not supplied as request fields.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] Delete Skill
>
> **Slug:** `PERPLEXITYAI_DELETE_SKILL`
>
> Delete a skill only if its current revision matches expected\_revision. Obtain the skill identifier and current revision from create or list skills. A successful deletion returns an empty acknowledgement. A revision conflict returns HTTP 409; retrieve the current revision before deciding to retry.
>
> #### Input Parameters
>
> **skill\_id** `string` *required*
>
> Skill identifier returned by create or GET /v1/skills.
>
> **expected\_revision** `string` *required*
>
> Opaque current revision token. Deletion fails with HTTP 409 if the skill has a newer revision, preventing a race with concurrent updates.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] Download Agent Response File
>
> **Slug:** `PERPLEXITYAI_DOWNLOAD_AGENT_RESPONSE_FILE`
>
> Download a file delivered by an Agent response's share\_file tool. Use the response ID and a file ID obtained from its file list. Returns a downloadable file rather than JSON-decoded or text-decoded provider bytes.
>
> #### Input Parameters
>
> **id** `string` *required*
>
> Agent response ID (resp\_\<...>), not a file ID.
>
> **file\_id** `string` *required*
>
> File ID from the agent response file list endpoint.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] Evaluate Decisions
>
> **Slug:** `PERPLEXITYAI_EVALUATE_DECISIONS`
>
> Evaluate named yes/no, choice, and score questions against shared content. Returns typed answers, probabilities, confidence where applicable, and token usage. The state plus any single question must fit within 256K input tokens; the complete body must be under 32 MiB. This is a non-streaming operation.
>
> #### Input Parameters
>
> **model** `string` *required*
>
> Decision model, e.g. pplx-decider-v1.2, pplx-decider-flash-v1.2, pplx-decider-latest, or pplx-decider-flash-latest. Provider also accepts perplexity/-prefixed latest aliases and legacy pplx-decider-v1.1-27b / pplx-decider-v1-27b. Unknown models return 400.
>
> **state** `string` *required*
>
> Shared content as string, object, or array. May contain OpenAI-style image\_url parts with base64 PNG, JPEG, or WebP data URLs; up to 8 images per question.
>
> **questions** `object` *required*
>
> 1–128 questions keyed by non-empty names. Mix noul, choice, and score questions; answers use the same names.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] Execute Agent
>
> **Slug:** `PERPLEXITYAI_EXECUTE_AGENT`
>
> Create Agent Response (POST /v1/agent). Orchestrates multi-step agentic workflows with built-in tools (web search, URL fetching, function calling, finance search, image search), reasoning, and multi-model support. Streaming is not supported by this action. At least one of 'model', 'models', or 'preset' must be provided. Available presets: 'fast-search', 'pro-search', 'deep-research'. The 'deep-research' preset generates very long responses (10K-100K+ words) with exhaustive multi-source analysis. Available models include Perplexity Sonar, OpenAI, Anthropic, Google, xAI, and NVIDIA models at direct provider rates. Use the List Models action to see available model identifiers.
>
> #### Input Parameters
>
> **input** `string` *required*
>
> The user's input/query to process.
>
> **model** `string`
>
> Model identifier (e.g. 'openai/gpt-5.2', 'perplexity/sonar'). Provider/model format. At least one of model, models, or preset is required.
>
> **tools** `array`
>
> Tools available to the agent: web\_search, fetch\_url, function, finance\_search, image\_search.
>
> **models** `array`
>
> Fallback chain of models (min 1, max 5). API tries each in order if prior fails. At least one of model, models, or preset is required.
>
> **preset** `string`
>
> Preset configuration: 'fast-search', 'pro-search', 'deep-research'. 'deep-research' generates very long responses (10K-100K+ words). At least one of model, models, or preset is required.
>
> **max\_steps** `integer`
>
> Maximum number of agent steps (1-10).
>
> **reasoning** `object`
>
> Configuration for reasoning behavior.
>
> **instructions** `string`
>
> System instructions to guide the agent's behavior.
>
> **response\_format** `object`
>
> Response format for structured output.
>
> **max\_output\_tokens** `integer`
>
> Maximum output tokens.
>
> **language\_preference** `string`
>
> Preferred response language as ISO 639-1 code.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] Get Agent Response
>
> **Slug:** `PERPLEXITYAI_GET_AGENT_RESPONSE`
>
> Retrieve a stored Agent response as a JSON snapshot (GET /v1/agent/\{id}). Poll this action for background responses until their status is terminal. Only responses created with store omitted or true can be retrieved. An unknown ID, a response owned by another account, or store=false produces HTTP 404.
>
> #### Input Parameters
>
> **id** `string` *required*
>
> Response ID returned by the Agent API (resp\_\<...>).
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] Get Async Chat Completion
>
> **Slug:** `PERPLEXITYAI_GET_ASYNC_CHAT_COMPLETION`
>
> Get Async Chat Completion (GET /v1/async/sonar/\{id}). Retrieves the result of an asynchronous chat completion request by its ID. Use this to poll for the result after creating an async job. The response includes the status and, when completed, the full completion.
>
> #### Input Parameters
>
> **request\_id** `string` *required*
>
> The unique identifier of the async chat completion request to retrieve.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] Get Member Usage Analytics
>
> **Slug:** `PERPLEXITYAI_GET_MEMBER_USAGE_ANALYTICS`
>
> Get organization Perplexity usage per member in daily UTC buckets. Requires a separate organization analytics API key (the same key as the v1 analytics API), not a standard inference API key. The Analytics API must be enabled for the organization. credit\_usage includes Model and Credit Source; query\_volume includes Feature, Project and Comet. Only complete UTC days are returned, never the in-progress day. Recent days may be absent while the analytics store syncs. Follow next\_page with the same query to retrieve more member rows; the requested window may span at most 90 days.
>
> #### Input Parameters
>
> **page** `string`
>
> Opaque next\_page cursor from a previous response. Reuse with otherwise identical query parameters.
>
> **limit** `integer`
>
> Member rows per page; provider default 50, maximum 100. A page can span multiple days.
>
> **dataset** `string ("credit_usage" | "query_volume")` *required*
>
> Dataset to query. query\_volume includes Feature, Project and Comet breakdowns, but not Model or Model Family.
>
> **end\_time** `integer`
>
> Exclusive UTC Unix seconds. Defaults to now; future values are capped at now. The window may span at most 90 days.
>
> **group\_by** `string` *required*
>
> Dimension to group daily usage by; only user\_email is supported.
>
> **start\_time** `integer` *required*
>
> Inclusive UTC Unix seconds. Snapped down to the day grid, so the first day can include earlier usage.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] Get Orchestrator Model
>
> **Slug:** `PERPLEXITYAI_GET_ORCHESTRATOR_MODEL`
>
> Retrieve one Orchestrator model by its full provider/model id. Returns model metadata and base input, output, and available cache token prices. A model unavailable on the Orchestrator API returns HTTP 404.
>
> #### Input Parameters
>
> **model** `string` *required*
>
> Full Orchestrator model id, including the provider prefix and slash, e.g. perplexity/kimi-k3. Supply the unencoded id.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] Get Skill
>
> **Slug:** `PERPLEXITYAI_GET_SKILL`
>
> Get a skill's metadata by identifier, optionally at a specific revision. Without revision, returns the active revision. The returned latest\_revision identifies the current latest revision even when an older revision is requested.
>
> #### Input Parameters
>
> **revision** `string`
>
> Opaque revision token; omit to retrieve the active revision.
>
> **skill\_id** `string` *required*
>
> Skill identifier returned by create or GET /v1/skills.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] Get Skill Download URL
>
> **Slug:** `PERPLEXITYAI_GET_SKILL_DOWNLOAD_URL`
>
> Get a short-lived URL to download a skill bundle as a ZIP, optionally for a specific revision; omit the revision to select the active revision. The response may also include revision and name metadata, and the bundle is not fetched.
>
> #### Input Parameters
>
> **revision** `string`
>
> Opaque revision token; omit to download the active revision.
>
> **skill\_id** `string` *required*
>
> Skill identifier returned by create or GET /v1/skills.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] Get Usage Analytics
>
> **Slug:** `PERPLEXITYAI_GET_USAGE_ANALYTICS`
>
> Get an organization's Perplexity Computer usage as a bucketed time series. Requires a separate organization analytics API key, generated by an org admin in Perplexity Settings → Organization → Computer, rather than a standard API inference key. Supply that key through the existing API-key authentication. Only complete UTC-grid buckets are returned; the in-progress bucket is omitted. Counts represent credits for credit\_usage, queries for query\_volume, distinct active members for daily\_active\_users, and threads for other datasets. Daily active-user counts are not additive across days. Analytics must be enabled for the organization (otherwise HTTP 404); rate limits apply across its keys.
>
> #### Input Parameters
>
> **page** `string`
>
> Opaque next\_page cursor from a previous response, valid only with the same query parameters.
>
> **limit** `integer`
>
> Buckets per page. Daily buckets: default 7, maximum 31. Hourly buckets: default 24, maximum 168.
>
> **dataset** `string ("credit_usage" | "connectors" | "artifacts" | "skills" | "spaces" | "workflows" | "task_durations" | "query_volume" | "daily_active_users")` *required*
>
> Dataset to query. query\_volume and daily\_active\_users are daily organization-level datasets and reject hourly buckets and user\_email.
>
> **end\_time** `integer`
>
> Exclusive window end in Unix seconds (UTC). Defaults to now; future values are capped at now. Must follow start\_time and span at most 90 days; the provider validates the effective window.
>
> **start\_time** `integer` *required*
>
> Inclusive window start in Unix seconds (UTC). Snaps down to the bucket grid, so the first bucket may include earlier usage. Must precede end\_time; the provider enforces a maximum 90-day window.
>
> **user\_email** `string`
>
> Filter to one current organization member's email. A nonmember returns a generic 400. Unavailable for query\_volume and daily\_active\_users; use the v2 endpoint for per-member query volume.
>
> **bucket\_width** `string ("1d" | "1h")`
>
> UTC-grid bucket size; defaults to 1d. query\_volume and daily\_active\_users do not accept 1h.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] List Agent Response Files
>
> **Slug:** `PERPLEXITYAI_LIST_AGENT_RESPONSE_FILES`
>
> List files an agent response delivered from its sandbox using share\_file. Only the authenticated account's files are visible. Returns file metadata, not inline content; use the file content endpoint to download a file. An unknown response id or one owned by another account returns HTTP 404.
>
> #### Input Parameters
>
> **id** `string` *required*
>
> Agent response identifier (resp\_\<...>), not a file id.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] List Async Chat Completions
>
> **Slug:** `PERPLEXITYAI_LIST_ASYNC_CHAT_COMPLETIONS`
>
> List a page of legacy asynchronous Sonar chat completion requests for the authenticated user, including their statuses. The response may include a next-page token, but this action does not accept one.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] List Models
>
> **Slug:** `PERPLEXITYAI_LIST_MODELS`
>
> List Models (GET /v1/models). Lists models available for the Agent API. Returns model identifiers that can be used with the Agent endpoint. The response follows the OpenAI List Models format for compatibility. This is a public endpoint that does not require authentication.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] List Orchestrator Models
>
> **Slug:** `PERPLEXITYAI_LIST_ORCHESTRATOR_MODELS`
>
> List available Orchestrator API models sorted by id, with base token prices. Requires a Perplexity API key. Prices are in USD per 1M tokens and include input, output, and available cache-write and effective cached-input rates. Use the returned full model ids with the Orchestrator API.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] List Skill Revisions
>
> **Slug:** `PERPLEXITYAI_LIST_SKILL_REVISIONS`
>
> List a skill's revisions, newest first. Supply a skill identifier returned by create or List Skills. Returns revision tokens and available name and creation metadata. Pass next\_page\_token as page\_token to retrieve another page.
>
> #### Input Parameters
>
> **limit** `integer`
>
> Maximum revisions to return, from 1 to 200; the provider defaults to 50.
>
> **skill\_id** `string` *required*
>
> Skill identifier returned by create or List Skills.
>
> **page\_token** `string`
>
> Opaque token from a previous response's next\_page\_token; omit for the first page.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] List Skills
>
> **Slug:** `PERPLEXITYAI_LIST_SKILLS`
>
> List skills in the authenticated Project, newest first. Returns skill identifiers and available name, revision, and timestamp metadata. To retrieve another page, pass the returned next\_page\_token as page\_token.
>
> #### Input Parameters
>
> **limit** `integer`
>
> Maximum skills to return, from 1 to 200; the provider defaults to 50.
>
> **page\_token** `string`
>
> Opaque token from a previous response's next\_page\_token; omit for the first page.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] Perplexity Search (Raw Results)
>
> **Slug:** `PERPLEXITYAI_SEARCH`
>
> Search the Web (POST /search). Returns raw, ranked web search results directly from Perplexity's index without LLM processing. Faster and cheaper than chat completions when you need raw results. Supports filtering by domain, date, language, country, and recency. Max 20 results per request. Important: search\_recency\_filter and date filters (search\_after\_date\_filter, search\_before\_date\_filter, last\_updated\_after\_filter, last\_updated\_before\_filter) are mutually exclusive. Use one or the other, not both.
>
> #### Input Parameters
>
> **query** `string` *required*
>
> Search query. Can be a single string or list of strings for multi-query search.
>
> **country** `string`
>
> ISO 3166-1 alpha-2 country code (e.g. 'US', 'GB', 'DE').
>
> **max\_tokens** `integer`
>
> Maximum total tokens of webpage content returned (1-1000000).
>
> **max\_results** `integer`
>
> Maximum number of search results to return (1-20).
>
> **max\_tokens\_per\_page** `integer`
>
> Maximum tokens retrieved from each webpage (1-1000000).
>
> **search\_domain\_filter** `array`
>
> Domains to restrict search results to (max 20).
>
> **search\_recency\_filter** `string ("hour" | "day" | "week" | "month" | "year")`
>
> Filter for search recency.
>
> **search\_language\_filter** `array`
>
> ISO 639-1 language codes to filter results (max 20, 2 chars each).
>
> **search\_after\_date\_filter** `string`
>
> Only include results published after this date (MM/DD/YYYY). Cannot be used with search\_recency\_filter.
>
> **last\_updated\_after\_filter** `string`
>
> Only include results last updated after this date (MM/DD/YYYY). Cannot be used with search\_recency\_filter.
>
> **search\_before\_date\_filter** `string`
>
> Only include results published before this date (MM/DD/YYYY). Cannot be used with search\_recency\_filter.
>
> **last\_updated\_before\_filter** `string`
>
> Only include results last updated before this date (MM/DD/YYYY). Cannot be used with search\_recency\_filter.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not

> [accordion] Update Skill
>
> **Slug:** `PERPLEXITYAI_UPDATE_SKILL`
>
> Upload a new ZIP bundle as a revision of an existing Project-owned skill. Supply the current revision as expected\_revision; a concurrent change causes HTTP 409 rather than overwriting it. Name and description are re-read from SKILL.md in the new bundle. Returns the updated skill and revision metadata.
>
> #### Input Parameters
>
> **file** `object` *required*
>
> New ZIP skill bundle containing SKILL.md. Name and description are re-read from the bundle.
>
> **skill\_id** `string` *required*
>
> Skill identifier returned by create or listing skills.
>
> **expected\_revision** `string` *required*
>
> Opaque revision token the skill must currently be at. The update fails with HTTP 409 if the skill has moved past it, preventing concurrent-write races.
>
> #### Output
>
> **data** `string` *required*
>
> Data from the action execution
>
> **error** `string`
>
> Error if any occurred during the execution of the action
>
> **successful** `boolean` *required*
>
> Whether or not the action execution was successful or not
