Perplexity AI provides conversational AI models for generating human-like text responses
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.
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.
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.
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.
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.
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.
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.
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).
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.
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.
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.
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.
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.
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.
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.
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.
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.
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.
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.
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.
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.
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.
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.
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.
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.
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.
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.
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.
The Perplexity AI integration connects your Dench AI CRM directly to Perplexity AI, so agents can read and act on your Perplexity AI data as part of everyday work — answering questions in chat, keeping your CRM in sync, and running automations without anyone copying data between tools.
28 actions are available for agents to invoke on your behalf. Every call runs through Perplexity AI's own authorization, scoped to the account you connect.
Sign in to your Dench workspace and open Integrations.
Find Perplexity AI and click Connect — you'll authorize access through Perplexity AI's own sign-in flow. No API keys or code required.
Ask an agent to use Perplexity AI in chat, or call it from an automation.
Manage or disconnect the connection any time from workspace settings.
The Dench Perplexity AI integration connects your AI CRM to Perplexity AI, so AI agents can work with your Perplexity AI data as part of chats, automations, and CRM workflows. You connect your account once, and every agent in your workspace can use it — governed by your workspace permissions.
The Perplexity AI integration currently exposes 28 actions, including Cancel Agent Response, Create Async Chat Completion, Create Chat Completion, Create Contextualized Embeddings, Create Orchestrator Chat Completion, and Create Orchestrator Message. Agents invoke them on your behalf from chat or from automations.
No. You connect Perplexity AI from your Dench workspace using Perplexity AI's own sign-in and authorization flow — no API keys to copy, no glue code to maintain.
Connections are authorized through Perplexity AI's own authentication flow, and Dench stores only the authorization needed to act on your behalf. You can review and disconnect the Perplexity AI connection from your workspace settings at any time.