Menu
Mattering Automations connector

A Perplexity MCP where every answer arrives with its sources

Connect one API key and search-grounded answers, raw web search, asynchronous deep research, and embeddings become workflow steps for visual automations and AI agents. Every response reports what it cost in USD, and the connection test runs on the one free authenticated endpoint, so verifying a key costs nothing.

8 ready-made nodes One credential to connect Works with Claude, ChatGPT, and Matter

What the Perplexity MCP connector does

The Perplexity connector for Mattering Automations is 8 nodes across the sonar API surface: chat completions answered with live web search, an asynchronous create and poll pair built for deep research runs that take minutes, the raw Search API that returns ranked results with extracted page content, plain and contextualized embeddings, and free listing nodes for async jobs and the Agent API model catalog. Streaming is pinned off by design: every node returns one JSON response per execution, which is exactly what a canvas step can consume.

Every node speaks MCP, the open protocol for connecting AI assistants to tools, so Claude, ChatGPT, and any MCP-capable assistant can use them, alongside Matter, the AI assistant built into the platform. An assistant can ask a question grounded in the live web and hand back not just an answer but the sources behind it, through the same nodes your workflows use.

Because Mattering is a work platform rather than middleware, research lands somewhere useful by default: a question sitting on a task in Mattering Productivity can come back as a cited answer in the task's comments, and a scheduled sweep can keep a Mattering database current with sourced answers instead of unverifiable model output.

Two things set this connector apart from a plain LLM node. First, answers are auditable: alongside the answer text, every chat response carries the citation URLs and per-source search results with title, url, date, and snippet, so a person or a downstream step can check where a claim came from. Second, spend is visible: Perplexity's API has no free tier, and every response includes a usage.cost breakdown in USD, so a workflow can log exactly what each run cost.

One connection, three surfaces

  • Visual automations. Drag Perplexity nodes onto the canvas and wire them to triggers, conditions, and other apps.
  • AI agents, yours or ours. These nodes speak MCP, the open protocol, so Claude, ChatGPT, and any MCP-capable assistant can use them, and Matter, the agent built into the platform, uses them natively.
  • The whole platform. Combine with native tasks, projects, and other connectors, no middleware in between.

Every Perplexity node, ready to drop on the canvas

8 nodes, generated from the connector itself so this list always matches what ships. Expand a category to browse.

Read and query nodes 4

Perplexity: Get Async Chat Completion

Fetch one asynchronous chat completion job by id. While data.status is CREATED or IN_PROGRESS the response field is empty; once COMPLETED, data.response holds the full completion (answer, citations, search_results, usage with cost), and on FAILED data.error_message says why. Free to call; poll this after the create node.

Perplexity: List Async Chat Completions

List your recent asynchronous chat completion jobs with their status and timestamps at data.requests. Free to call and read-only, which makes it the connector's credential health check: a bad key gets a 401, a good key gets a 200 at no cost. The endpoint takes no parameters, so the config panel is intentionally empty.

Perplexity: List Models

List the model catalog for Perplexity's Agent API in the OpenAI list format (ids like anthropic/claude-sonnet-4-6 or openai/gpt-5.5). Honest scope note: this does NOT list the sonar chat models - those are the fixed sonar family named in the chat node's model field. Free; the endpoint requires no authentication, so it also cannot verify your key. The endpoint takes no parameters, so the config panel is intentionally empty.

Perplexity: Search the Web

Run Perplexity's raw web search and get ranked results with extracted page content, without invoking a model. Each item in data.results carries title, url, snippet, date, and last_updated. Read-only but PAID: billed at a flat per-request price, which is why its Test Class is not Safe Read. Path note: this endpoint is /search with no /v1 prefix.

Action nodes 4

Perplexity: Chat Completion (Sonar)

Ask a Perplexity sonar model a question answered with live web search. The response carries the answer at data.choices[0].message.content, the source URLs at data.citations, rich per-source detail (title, url, snippet, dates) at data.search_results, and token usage with a USD cost breakdown at data.usage. Billed per token plus a per-request search fee tiered by search context size. Streaming is always off: the node returns one JSON response.

Perplexity: Create Async Chat Completion

Submit a sonar chat completion as an asynchronous job and get back a request id immediately, instead of holding the connection open. Built for sonar-deep-research runs that take minutes. Poll the id with the Get Async Chat Completion node until data.status is COMPLETED, then read the full completion from data.response. Billing is identical to the synchronous node.

Perplexity: Create Contextualized Embeddings

Embed document chunks so each chunk's vector also encodes its document-level context, improving RAG retrieval over plain chunk embeddings. Send a JSON array of documents, each a list of chunk strings, and get the same nested shape back at data.data[doc].data[chunk].embedding (base64-encoded, like the plain embeddings node). Billed per input token.

Perplexity: Create Embeddings

Turn text into embedding vectors with a Perplexity pplx-embed model. Send one string or a JSON array of strings, get one vector per input at data.data[i].embedding. Note the vectors come back BASE64-ENCODED (int8 or packed binary), not as plain JSON float arrays; downstream consumers must decode them. Billed per input token, reported at data.usage.

Connecting Perplexity takes one credential

Perplexity issues API keys from the API Keys page in its console at console.perplexity.ai. Mattering sends the key as a bearer credential on every request. The API is pay as you go with no free tier or trial credits; billing is per token plus a tiered per-request search fee on chat, a flat per-request price on the Search API, and per token on embeddings, with every response reporting its own cost in USD.

1

Create an API key on the API Keys page in the Perplexity console (console.perplexity.ai).

2

In Mattering Automations, open Connectors, click Add Connector, choose Perplexity, and create a credential with that key.

3

Click Test Connection. Mattering lists your asynchronous chat jobs, the one authenticated Perplexity endpoint that is free to call, so a bad key fails with a 401 and a good key passes at no cost.

4

All 8 Perplexity nodes appear in the canvas palette. The two async reads and the model list are free to call; every other node bills from its first run.

Common uses for the Perplexity MCP

Answers a person can audit

Perplexity: Chat Completion (Sonar) asks a sonar model a question answered with live web search. The response carries the answer, the plain citation URLs, and per-source search results with title, url, date, and snippet, so a reviewer or a downstream step can check every claim against where it came from.

Deep research that runs as a job

Perplexity: Create Async Chat Completion submits a chat request, sonar-deep-research included, and returns a request id immediately instead of holding a connection open for minutes. Perplexity: Get Async Chat Completion polls that id, free of charge, until the status reaches COMPLETED and the full report arrives with its citations, search results, and cost.

Raw web search with the page content included

Perplexity: Search the Web runs Perplexity's search directly, no model involved, and returns ranked results where each item carries title, url, snippet, dates, and extracted page content. It is read-only on the web but a paid call at a flat per-request price, and the connector says so rather than pretending it is free.

Embeddings, plain and contextualized

Perplexity: Create Embeddings turns text into vectors with the pplx-embed models, and Perplexity: Create Contextualized Embeddings embeds document chunks so each vector also carries its document-level context, the fix for RAG's isolated-chunk blind spot. Vectors return base64 encoded, which keeps workflow payloads small; decode them downstream.

Spend tracking built into every response

Every response includes a usage.cost object with a USD breakdown: input and output token costs, the per-request search fee, and on deep research runs the citation, reasoning, and search query costs, summed into a total. A workflow can append that total to a Mattering database so the team sees what each answer cost.

New models without a connector update

The model on every node is a plain config field, not a locked list, so new Perplexity models work the day they ship. The current chat family is sonar, sonar-pro, sonar-reasoning-pro, and sonar-deep-research, with sonar as the default; the embedding nodes default to pplx-embed-v1-0.6b and pplx-embed-context-v1-0.6b.

Example: A research question on a task becomes a cited answer in its comments

Built on the Automations canvas in a few minutes, or described in plain language to Matter and built for you.

1

Trigger: a task in Mattering Productivity is tagged for research, with the question in its description.

2

Perplexity: Chat Completion (Sonar) asks the question, answered with live web search on the sonar model.

3

The response carries the answer, the citation URLs, and per-source search results with title, url, date, and snippet.

4

The answer and its sources are formatted into a comment that lands back on the task, so review happens next to the evidence.

5

The usage.cost total from the same response is appended to a Mattering database, so the team sees research spend per question.

6

Questions flagged for a full report route to Perplexity: Create Async Chat Completion on sonar-deep-research instead, with Perplexity: Get Async Chat Completion polling until the report is ready.

Perplexity MCP questions, answered

What Perplexity operations are supported?

8 nodes on the sonar API surface: chat completions with live web search, the asynchronous create, poll, and list trio for long runs, the raw Search API, plain and contextualized embeddings, and the Agent API model catalog listing. Deliberately not shipped: the Agent API itself (an agentic loop with third-party model passthrough and binary file downloads, which does not fit a canvas node shape; the sonar surface remains live and documented), Computer usage analytics, and the separate Gateway API for routing to third-party models.

Does this work with Claude, ChatGPT, or other AI assistants?

Yes. Mattering exposes these nodes over MCP, an open protocol for connecting AI assistants to tools, so any MCP-capable assistant, including Claude and ChatGPT, can connect to your workspace and use them. Matter, the assistant built into the platform, uses the same nodes natively.

How is this different from a plain LLM node?

The sonar models search the live web before answering and return their sources. Alongside the answer text, the response carries citation URLs and per-source search results with title, url, date, and snippet, so answers are auditable rather than taken on faith. For questions about anything current, that is the reason to route an automation through Perplexity instead of a model working from training data alone.

What does it cost?

Perplexity's API is pay as you go with no free tier or trial credits. The mechanisms: chat completions bill per input and output token plus a per-request search fee tiered by search context size, and deep research runs add citation tokens, reasoning tokens, and per-search fees; the Search API bills a flat price per request; embeddings bill per input token. The connector never invents prices, so check Perplexity's pricing page for current numbers; every response includes a usage.cost breakdown in USD, so actual spend is visible per run.

Why is the search node treated as a paid write for testing when it only reads?

Honesty about billing. Perplexity: Search the Web has no side effects, but it bills a flat price per request, so classing it as a safe read would let test batches and casual probes spend money silently. It sits in the write tier for testing purely so nothing treats it as a free call. The genuinely free reads are the async job list, the async job fetch, and the model list.

Which models can I use, and what does the model list node actually list?

The model field on every node is free text with a documented default, so new models work without a connector update. The chat family today is sonar, sonar-pro, sonar-reasoning-pro, and sonar-deep-research; the embedding nodes default to pplx-embed-v1-0.6b and pplx-embed-context-v1-0.6b. One honest wrinkle: Perplexity: List Models returns the catalog for Perplexity's separate Agent API, provider-prefixed ids for third-party models, not the sonar chat family, and its endpoint requires no authentication, so it cannot verify a key either.

How do long deep research runs work?

As an asynchronous job. Perplexity: Create Async Chat Completion submits the full chat request and returns a request id with status CREATED; Perplexity: Get Async Chat Completion polls the id, free of charge, until the status reaches COMPLETED and carries the full report, or FAILED with an error message. The synchronous chat node can run deep research too, but it carries a 300 second timeout, so the async pair is the right shape for reports that take minutes. Billing is identical either way.

Can responses stream?

No, by design. The connector pins streaming off on the synchronous chat node and inside the async request, and workflow config cannot re-enable it. A canvas step consumes one JSON response per execution, so server-sent events would have nothing to attach to; the node simply returns the finished answer with its citations, search results, and cost.

8 nodes waiting on the canvas

Put Perplexity to work inside your workspace

Connect once, then use every node in visual automations and AI agent workflows across the Mattering platform.

All product names, logos, and brands are property of their respective owners and are used for identification purposes only. Use of these names, logos, and brands does not imply endorsement or affiliation. Perplexity is a product of its respective owner; Mattering is an independent platform.

Start Free Explore our Products