Pricing

Start free. Pay for what your product thinks.

Connect your docs and test for free. When Pernoex answers, it spends credits, and we translate that into the only number you care about: conversations a month. 7-day free trial on every plan.

How it works

Docs to cited answers in three steps.

01

Connect your docs

Crawl your site, upload PDFs, or connect Confluence and GitHub. Indexing and testing are always free, invite your team, ask real questions.

✻ Free · unlimited
02

It thinks, it cites

Every answer is grounded in your content with a source you can open. Credits are spent only when it reasons, a chat reply, an inline explanation, an IDE query.

✻ Only pay when it answers
03

Measured in conversations

Each plan includes a monthly credit pool we express as estimated conversations. Need more? Add a credit pack in one click, never throttled mid-chat.

✻ Change plans anytime
What's a credit?

You think in conversations. We handle the rest.

Credits are consumed when your product thinks, and the cost scales with the complexity of the reasoning, not a flat per-message fee. A quick lookup is cheap; a multi-source answer with a live API call costs more.

So we don't make you do the maths. Every plan shows an honest estimated conversations / month, and a credit pack is always one click away.

Credits are spent when it reasons.
Chat & inline answerLower
Multi-source / IDE queryHigher
Answer + Custom ActionHighest
Connecting & testingFree
Run low?+ pack, 1 click
Plans

Priced to start small, built to scale.

Starter
Start free
Most popular
Growth
Start free

Enterprise

Custom credit volume, SSO, SLA guarantees, data residency and dedicated support, for teams deploying across every market.

Contact sales

Conversation estimates are typical ranges, not caps; actual usage depends on answer complexity. Extra credits top up in one click. Annual billing saves two months. An L1 support agent costs $48–80K/year fully loaded¹, Pernoex planning benchmark, 2026.

Compare plans

Every plan cites every answer.

The difference is volume, surfaces and control, not whether the answers are grounded. They always are.

CapabilityStarterGrowthScale
Credits & Limits
Monthly credits
Est. conversations
Active projects
Team seats
Allowed domains
Overage rate (/1k credits)LoadingLoadingLoading
Extra project add-onAvailableAvailable
Extra seat add-onAvailableAvailable
Credit pack add-ons
Surfaces
The Console
Embedded
Inline Explainer
Voice support (add-on)LoadingLoading
Voice Add-on
Voice add-on priceLoadingLoading
BYOC Voice (bring your own carrier)AvailableAvailable
Included minutesLoadingLoading
Phone numbers includedLoadingLoading
Overage rate (geographic)LoadingLoading
Overage rate (toll-free)LoadingLoading
Toll-free minute multiplierLoadingLoading
Extra phone number add-on
Minute pack add-ons
Knowledge
Document uploads
Web crawling
PDF upload
Multi-language (50+)
Data sources (Confluence, GitHub)
Scheduled re-crawlsDaily / Weekly
Developer
MCP access (IDE integration)
Configurable MCP usage cap
Server-side REST API
Webhooks
Integrations
Slack integration
Custom Slack bot
Custom Actions
Conversation handoff (Front · Intercom)
Analytics
Usage dashboard
Traffic sources
Content gap detection
Business impact dashboard
Analytics export (CSV)
Branding & Compliance
White-label (remove badge)Add-onIncludedIncluded
Activity audit log
Role-based access control
Multi-company support
Support
Email support
Priority support
Dedicated account manager
What you unlock as you grow

It doesn't just answer. It acts.

Growth

Custom Actions

Connect your APIs so the AI can check an order, reset a password or look up an account mid-conversation, on chat, Slack and voice, every action audited.

Growth

Content-gap detection

See the questions your docs can't answer yet, with AI-suggested drafts to fill them, so the knowledge base compounds.

Scale

Human handoff, full context

When the AI shouldn't answer, it transfers to your team in Front or Intercom with the full thread, user identity and page, no one starts from scratch.

Questions

Before you get started.

How do credits work?

Credits are consumed when your product thinks, a chat reply, a highlight explanation, an IDE query. Cost scales with the reasoning, so we show you the plain-English number: conversations per month. Top up in one click anytime.

What happens when it can't answer?

It asks a clarifying question rather than guessing, flags the gap so you can fill it, and, on Scale, hands off to a human with full context. Pernoex never fabricates a confident reply.

Is the trial really free?

Yes, 7 days on any plan, cancel anytime, card required to start. Go live in minutes with one script tag; connecting and testing stay free after that, too.

How does voice pricing work?

Voice is a Growth or Scale add-on with phone numbers in 60+ countries. Same knowledge base, same citations, billed on top of your plan.

Your product already has the answers. Give it a voice.

One script tag, under five minutes. Free for 7 days. Cited by default.