# Perplexity — AI answer engine with an agentic Computer, the Comet browser, and a developer API for web-grounded search.

> Source: The Agents Index — https://theagentsindex.com/perplexity (structured, researched, re-verified)
> Facts last verified: 2026-09-02

Perplexity is an AI answer engine that responds to questions with inline citations drawn from live web search. It also ships Perplexity Computer for multi-step agentic work, the Comet browser, and an API platform for developers.

| Fact | Value |
| --- | --- |
| Website | https://www.perplexity.ai |
| Price | from $20/month billed monthly, $17/month billed annually |
| Free tier | Yes |

Free plan at $0/month: 'Good for limited daily usage', citations in every answer, access to basic AI models; web answers, file uploads and asset generation are marked 'Limited', and Computer credits, connectors and Model Council are absent.

## What it does

**Writes code** · Yes — Perplexity Computer builds apps and sites from a natural-language prompt, and the Agent API sandbox has the model write the code itself.

> Create production ready apps, websites, reports, and more.
> — https://www.perplexity.ai/products/computer · Computer product page, 'What Computer does for you' section, item 04 'Create and build'

**Debugs and fixes** · Yes — Computer is sold as fixing code from a plain-language instruction; the sandbox docs list reproducing an error as a use case.

> Analyze data, create content, build sites, fix code, and more just by telling Computer what you need in plain language.
> — https://www.perplexity.ai/enterprise · Enterprise page, 'Tackle complex projects' card

**Writes tests** · No — The enumerated GitHub-connector actions cover cloning, changing, committing and opening a pull request; test generation appears nowhere in that list or in the sandbox use cases.

> the agent can clone private repositories, inspect and change code, commit changes, push a branch, and create or update a pull request
> — https://docs.perplexity.ai/docs/agent-api/tools/connectors · Connectors, GitHub connector section, capability sentence

**Reviews code** · Partly — Review happens through the managed GitHub connector, which the same note marks as preview; there is no dedicated diff-review product.

> The GitHub connector lets the model run git and gh in the sandbox to review code, make changes, and open pull requests.
> — https://docs.perplexity.ai/docs/agent-api/tools/mcp · Overview, Connectors note (marked Preview)

**Whole-codebase work** · Partly — The agent clones a whole repository and works in it through git and gh, but Perplexity never claims a migration carried across a repo, and the connector is labelled preview.

> With both tools enabled, the agent can clone private repositories, inspect and change code, commit changes, push a branch, and create or update a pull request.
> — https://docs.perplexity.ai/docs/agent-api/tools/connectors · Connectors page, 'Use connectors in the sandbox' section, GitHub connector

**Ships it** · Partly — Through the API the agent pushes a branch and opens or updates a pull request; no deployment step or target is documented.

> With both tools enabled, the agent can clone private repositories, inspect and change code, commit changes, push a branch, and create or update a pull request.
> — https://docs.perplexity.ai/docs/agent-api/tools/connectors · Connectors page, 'Use connectors in the sandbox' section, GitHub connector

**Done-for-you service** · No — Self-serve subscription software across web, desktop, mobile, Slack and Microsoft 365; Perplexity sells seats and support, not a build-and-run engagement.

> Available to Pro and Max subscribers on desktop, mobile, Slack, and Microsoft 365.
> — https://www.perplexity.ai/products/computer · Computer product page, 'Put Computer to work' closing section

## How much it does unattended

**Runs autonomously** · Yes — Computer takes recurring tasks and executes them unattended in the background; runs are described as lasting hours or months.

> Set recurring tasks that Computer executes autonomously and persistently over time.
> — https://www.perplexity.ai/products/computer · Computer product page, Features item 01 'Background tasks and continuous monitoring'

**Multi-agent** · Yes — Computer fans a prompt out to subagents; the API's wide-research preset runs the same wide-and-deep pattern in the background.

> Prompt in natural language. Computer deploys subagents to handle the rest.
> — https://www.perplexity.ai/products/computer · Computer product page, 'How it works' step 01 'Give Computer tasks'

**Agent permissions** · Partly — API runs are bounded by the connected account's own permissions, an allowed_tools allowlist and a max_steps loop cap; no per-action human approval prompt is documented.

> The agent can access only repositories and perform only actions that your connected GitHub account permits.
> — https://docs.perplexity.ai/docs/agent-api/tools/connectors · Connectors page, GitHub connector subsection

## Which models it runs on

**Claude** · Yes — Claude is one of the models the consumer product orchestrates, and the Agent API exposes Opus, Sonnet and Haiku by model id.

> Perplexity orchestrates multiple top models and selects the best one per query, including ChatGPT, Google Gemini, Anthropic Claude, and NVIDIA Nemotron.
> — https://www.perplexity.ai/hub/pricing · Pricing page FAQ, 'Which AI models come with each plan?'

**GPT** · Yes — The Agent API lists the GPT-5 family with per-token rates; the consumer plans surface GPT as a selectable preferred model.

> GPT-5 family — flagship, mini, and nano variants.
> — https://docs.perplexity.ai/docs/agent-api/models · Agent API Models page, 'Available Models', OpenAI tab card

**Open models** · Yes — Open-weight Nemotron, DeepSeek, GLM and Kimi models are callable by id and hosted by Perplexity; the Router API is a further open-weight catalogue in private preview.

> Nemotron open-weight reasoning models.
> — https://docs.perplexity.ai/docs/agent-api/models · Agent API Models page, 'Available Models', NVIDIA tab card

**Model choice** · Yes — Paid consumer plans set a preferred model, and the API takes an explicit model id across multiple provider catalogues; otherwise Perplexity picks per query.

> Pro and Max both allow selecting a preferred model.
> — https://www.perplexity.ai/hub/pricing · Pricing page FAQ, 'Which AI models come with each plan?'

## Where you use it

**In your editor** · Partly — There is no Perplexity IDE extension; it enters the editor as an MCP tool that Claude Code, Cursor or VS Code calls.

> This page is for adding Perplexity as an MCP tool inside clients such as Claude Code, Cursor, VS Code, and Claude Desktop.
> — https://docs.perplexity.ai/docs/getting-started/integrations/mcp-server · Perplexity API MCP Server page, Overview note

**On the command line** · Yes — A first-party pplx binary runs web search and page-snippet extraction from a terminal and returns JSON for pipelines.

> The pplx CLI returns structured JSON from the Perplexity Search API, making it useful for shell pipelines, interactive terminal work, and coding agents that need current web results
> — https://docs.perplexity.ai/docs/cli/overview · Perplexity CLI page, opening paragraph

**In your pipeline** · Partly — The Agent API can open and update pull requests through the GitHub connector, which is in preview; no CI runner or GitHub Action is documented.

> With both tools enabled, the agent can clone private repositories, inspect and change code, commit changes, push a branch, and create or update a pull request.
> — https://docs.perplexity.ai/docs/agent-api/tools/connectors · Connectors page, 'Use connectors in the sandbox' section, GitHub connector

**In a browser** · Yes — The answer engine is a hosted web platform, with Comet as a downloadable browser and apps on desktop and mobile alongside it.

> Yes, the API offers exactly the same internet data access as Perplexity's web platform.
> — https://docs.perplexity.ai/docs/resources/faq · FAQ page, 'Search & Results' section, internet data access question

## Whose machine it runs on

**Self-hosted** · No — Perplexity runs the compute on its own AWS estate in North America; no self-hosted, on-prem or VPC deployment is documented.

> Our compute is hosted via Amazon Web Services in North America.
> — https://docs.perplexity.ai/docs/resources/faq · FAQ page, 'Data, Privacy & Security' section, 'Where Are Perplexity's Language Models Hosted?'

**Open source** · Partly — The MCP server is open source and installable as a local stdio server; the answer engine, Computer and Comet are closed.

> Local MCP server: run the open-source server on your machine over stdio.
> — https://docs.perplexity.ai/docs/getting-started/integrations/mcp-server · Perplexity API MCP Server page, Overview, two ways to connect

## What it costs to run

**How it meters** · Yes — Consumer and Enterprise plans bill per seat per month, Computer meters credits per seat, and the API bills per token plus a fee per tool invocation.

> the Agent API charges per token at direct provider rates with no markup, plus a per-invocation fee for tools; the Search API charges per request; and Embeddings charge per token.
> — https://docs.perplexity.ai/docs/resources/faq · FAQ page, 'Account, Billing & Limits' section, 'How Does API Pricing Work?'

**Free tier** · Yes — A permanent $0 plan gives cited answers on the top models with limited weekly usage; the API instead requires purchased credits.

> Free includes access to the top AI models with limited weekly usage.
> — https://www.perplexity.ai/hub/pricing · Pricing page FAQ, 'How is Pro different from Free?'

**API access** · Yes — A separate API Platform exposes Agent, Search, Router and Embeddings APIs with Python and TypeScript SDKs, billed on prepaid credits rather than a consumer plan.

> API keys are your credentials for accessing the Perplexity API.
> — https://docs.perplexity.ai/docs/getting-started/projects · Projects & Billing page, 'Managing API Keys' section

**MCP server** · Yes — Perplexity hosts a remote MCP server another agent can query with an API key, and publishes a local stdio server as an alternative.

> Remote MCP server (recommended): hosted by Perplexity at https://api.perplexity.ai/mcp. No installation, nothing to update, works with any client that supports remote MCP servers.
> — https://docs.perplexity.ai/docs/getting-started/integrations/mcp-server · Perplexity API MCP Server page, Overview, two ways to connect

**Bring your own key** · No — Perplexity fronts the provider accounts and bills tokens against a Perplexity key; no bring-your-own-key, own-inference or own-cloud path is documented.

> All models are accessed directly from first-party providers with transparent token-based pricing.
> — https://docs.perplexity.ai/docs/agent-api/models · Agent API Models page, 'Available Models' introduction

## Buying it for a team

**A company can buy it** · Yes — Enterprise Pro and Enterprise Max are sold to organisations, can be mixed within one org, and admins assign each member's tier.

> Yes. An organization can include both Enterprise Pro and Enterprise Max users. Admins set each member's tier under Settings, then Members.
> — https://www.perplexity.ai/hub/pricing?p=enterprise · Pricing page, Enterprise tab FAQ, 'Can one organization mix Enterprise Pro and Enterprise Max users?'

**Seat model** · Yes — Enterprise Pro is $34 and Enterprise Max $271 per seat per month billed annually; no seat floor or ceiling is published.

> $34 / month per seat when billed annually
> — https://www.perplexity.ai/hub/pricing?p=enterprise · Pricing page, Enterprise tab, Enterprise Pro card price cell

**Pooled budget** · No — Computer credits are allotted per seat, 500 a month on Enterprise Pro and 15,000 on Enterprise Max, not drawn from one shared org balance.

> 500 monthly Computer credits/seat
> — https://www.perplexity.ai/hub/pricing?p=enterprise · Pricing page, Enterprise tab, Enterprise Pro card feature list

**Admin controls** · Partly — User management arrives with Enterprise Pro; role-based access control and custom roles sit in the custom-priced governance tier and the comparison table marks that row 'Contract required'.

> Manage user permissions and integrate with SSO
> — https://www.perplexity.ai/hub/pricing?p=enterprise · Pricing page, Enterprise tab, Enterprise Pro card feature list

**Audit log** · Partly — Audit logs are named on the Enterprise Max card and in the custom governance tier; the comparison table marks the audit-log row 'Contract required'.

> Premium security features including SCIM, audit logs, and data retention settings
> — https://www.perplexity.ai/hub/pricing?p=enterprise · Pricing page, Enterprise tab, Enterprise Max card feature list

**Single sign-on** · Partly — SAML 2.0 and OIDC through your own IdP, but only for organisations with an active org subscription; the same page states it is not available on personal Pro or Max.

> The portal supports SAML 2.0 and OIDC connections and walks you through your specific IdP.
> — https://docs.perplexity.ai/docs/admin/signing-in · Signing In to the API Portal page, 'Configuring SSO for Your Organization', step 'Connect Your Identity Provider'

## What happens to your code

**Opt out of training** · Partly — Guaranteed on Enterprise Pro and Enterprise Max; the comparison row is blank for individual Pro, while the API is documented as zero-retention and never trained on.

> Guaranteed no training on your data
> — https://www.perplexity.ai/hub/pricing?p=enterprise · Pricing page, Enterprise tab, Enterprise Pro card feature list

**Data residency** · Partly — The jurisdiction is stated as AWS in North America, but no region or residency choice is offered, and enterprise data-retention configurability is contract-gated.

> Our compute is hosted via Amazon Web Services in North America.
> — https://docs.perplexity.ai/docs/resources/faq · FAQ page, 'Data, Privacy & Security' section, 'Where Are Perplexity's Language Models Hosted?'

**Getting out** · Partly — Portability is a request under the privacy notice rather than a self-serve export, and nothing published states whether unused Computer credits expire.

> In certain cases, you may ask to receive access to your data in a portable, machine-readable form.
> — https://www.perplexity.ai/hub/legal/privacy-notice · Privacy Notice, 'Your Rights and Choices', Access

**Certifications** · Yes — SOC 2 Type II, HIPAA, GDPR and PCI DSS are named on the Enterprise Pro card; the API docs add a 2025 HIPAA gap assessment and a CAIQlite assessment via the Trust Center.

> SOC 2 Type II, HIPAA, GDPR, PCI DSS compliance
> — https://www.perplexity.ai/hub/pricing?p=enterprise · Pricing page, Enterprise tab, Enterprise Pro card feature list

## Provenance

Every fact above comes from the vendor. Nothing here is independently corroborated yet.

## Coverage elsewhere

Pages about this product that the vendor does not control. The label states how independent each one is.

| Source | Kind | Read |
| --- | --- | --- |
| theverge.com | Independent source | 2026-08-23 |
| cnbc.com | Independent source | 2026-08-23 |
| techcrunch.com | Independent source | 2026-08-23 |
| stockanalysis.com | Independent source | 2026-08-23 |
| valueaddvc.com | Independent source | 2026-08-23 |
| pymnts.com | Independent source | 2026-08-23 |

## What changed

- 2026-09-02 — Capabilities · Commercial terms · Free tier · Price
- 2026-09-01 — Capabilities · Commercial terms · Price
- 2026-08-31 — Capabilities · Commercial terms · Free tier · Price
- 2026-08-22 — Alternatives · One-liner · Summary

## Correcting this record

Anything marked "Not established" is a question we have not answered — it is not a No. If you represent Perplexity, or anything here is wrong or out of date, you can supply it with a source: the operations this site accepts from an agent are listed at https://theagentsindex.com/.well-known/agents.json — that catalogue needs no key to read, and each entry states its own auth.

Supplying evidence is free. Nothing that can be bought here changes what a field says, whether a gap is closed, or how this record is scored.