$99 for your first mod, $75 for every mod after. Own production-ready source code forever—zero subscriptions, zero platform fees.
HelpfulMods

Service

Website Chat

AI support chat grounded in your docs, with staff replies in the same conversation.

Turn visitor questions into sales conversations with source-grounded AI chat. Add an embeddable sales and support widget that checks answers against your approved business information, shows source links and lets visitors leave a message for your team. Staff replies stay in the same saved conversation. Purchases, refunds and account changes remain under your app's control.

$99 first mod · $75 each additional mod · One-time license

See the included chat widget

View full-size screenshot

The built-in chat widget, shown with sample data.

Test-drive the live workspace with sample data. Nothing is charged or sent. Your sample workspace lasts up to 15 minutes.

Website Chat chat widget preview, version 0.1.8, using sample data.
Included in version 0.1.8 · Sample data

Example scenario

See it live: Mallow, the chat widget in the corner, is built on our Website Chat mod. Ask it anything about Helpful Mods.

Help a visitor choose, then keep the conversation

Add your plan details and setup guides. A visitor asks which plan fits, follows a source link and returns to the saved conversation. If they leave a message for your team, staff reply in the private inbox and the visitor reads it in the same chat.

From setup to the result

  1. Add your product facts, prices, policies and useful links, then choose a model budget.
  2. Run the Node.js service with persistent SQLite storage and embed the widget.
  3. Test your common customer questions, source links and saved conversations.
  4. Visitors leave messages for your team, and staff reply from a private web inbox.

What the package handles

What you can build with it

  • Business-specific answers with source references and links to your approved pages.
  • Saved conversations that survive a restart, with a visitor-controlled delete option.
  • Visitor message handoff, a private staff inbox and replies in the same chat thread.
  • An embeddable widget with an animated assistant, reduced-motion support and an image fallback.

Included capabilities

  • An isolated iframe widget limited to the websites you choose
  • Business facts updated through the protected console, local JSON or an approved HTTPS feed
  • Business-specific answers, saved conversations, source links and visitor message handoff
  • Saved conversations keep newly sent messages and their delivery state intact while history loads. The widget applies your configured theme to messages, input and visitor handoff.

Before you choose

Architecture & Runtime Specs

Who it is for

  • Businesses answering product, pricing and setup questions on their website.
  • Teams that want saved visitor conversations and a private inbox for follow-up.
  • Site owners running a small Node.js service alongside their website.

Hosting and connections

V0.1.9

Host this as a Node.js 24 service with a persistent SQLite disk and one process per database. Live answers require your OpenAI API account; approved facts, staff controls and source freshness checks run in the installed service.

Self-host on Node.js 24 or later and SQLite on persistent local storage with zero Helpful Mods platform or per-seat fees. Your customer data stays in your database.

Connects directly to your own OpenAI accounts with zero middleman fees.

Show the integration

Your app backendWebsite ChatOpenAI

Your app calls the mod on your server. The mod saves its records in your database and uses your own provider accounts for the features you enable.

Start the configured Website Chat service with persistent SQLite storage. Set MOD_ORIGIN to its configured backend origin. Use the configured HTTPS service origin in production.

Open a visitor conversation

Open a visitor conversation, then read its saved history. The service returns a token for that conversation. This probe reads its history, then deletes the sample through /api/forget. In your app, retain the token to resume the conversation and use the deletion route when the visitor chooses to forget it. An expired token requires a new session. Visitor tokens are separate from app sign-in and staff access.

Visitor session request example

Save this as integration-example.mjs and run node --env-file=.env integration-example.mjs. These requests do not send email, take payment or call an AI model.

const origin = process.env.MOD_ORIGIN;
if (!origin) throw new Error('Set MOD_ORIGIN to your chat service origin');
const started = await fetch(new URL('/api/session', origin), {
  method: 'POST', redirect: 'error', credentials: 'omit',
  signal: AbortSignal.timeout(10000), headers: { Origin: origin },
});
if (!started.ok) throw new Error('Session request failed: ' + started.status);
const { token } = await started.json();
const response = await fetch(new URL('/api/history', origin), {
  redirect: 'error', credentials: 'omit', signal: AbortSignal.timeout(10000),
  headers: { Origin: origin, Authorization: 'Bearer ' + token },
});
if (!response.ok) throw new Error('History request failed: ' + response.status);
console.log(await response.json());
const forgotten = await fetch(new URL('/api/forget', origin), {
  method: 'DELETE', redirect: 'error', credentials: 'omit',
  signal: AbortSignal.timeout(10000),
  headers: { Origin: origin, Authorization: 'Bearer ' + token },
});
if (!forgotten.ok) throw new Error('Cleanup request failed: ' + forgotten.status);

A useful handoff to your coding agent

How setup works

Point your coding agent at README.md and setup/SKILL.md. The zip includes architecture docs, an integration guide, an error reference, and automated checks so your agent can work directly from the source package. Live provider setup connects to your own accounts.

  1. Architecture scan. Your coding agent inspects your codebase, routes, database and authentication, then plans how to connect the mod.
  2. Answer a few quick questions. Your agent asks how you want the feature configured and connects your API keys in your private environment.
  3. Run the tests and go live. Your agent runs the included verification suite, checks the customer workflow and tests your connected services before launch.

What your agent will ask you during setup

  • Which products, prices, policies and links should the assistant know?
  • Where will the Node.js service and SQLite database run, and which websites can embed the widget?
  • Who updates the facts and checks the inbox, and what retention period and AI budget do you want?

What your agent checks

  • Test the widget and animations in local demo mode without calling OpenAI.
  • Check answers, source links and questions your business information does not cover.
  • Restart the service and confirm the conversation and original request are retained.
  • Leave a test message, reply from the staff inbox and use Check replies to read it in chat.
  • Confirm expired knowledge or an unresolved OpenAI failure pauses AI answers instead of producing an unconfirmed reply.

Focused Scope & Extensibility

Included scope, capacity and customization
  • Run one business and one process per persistent SQLite database on Node.js 24 or later.
  • Answers are checked against your approved business facts before display. Review real visitor questions and keep your knowledge base current to improve the answers.
  • Choose a daily AI budget from estimated usage and check actual charges in your OpenAI account.
  • Staff reply in the private web inbox and visitors use Check replies in the same chat. Email or push alerts can connect through your notification services.
  • Read-only answers and message intake leave purchases, refunds and account changes under your app's control.

Before you decide

Will this work with my existing app?

Yes. Website Chat runs alongside your existing website as a Node.js 24 service with SQLite on persistent local storage. Embed the included widget on your chosen domains and connect your OpenAI account for live answers. Run one process per database.

What will my agent need from me?

Access to your codebase, your answers to the setup questions above, and API keys for any connected services. Store keys in your own private .env environment or secret manager.

What happens if setup fails?

The included setup/error-reference.json gives your coding agent error explanations and recovery steps. Every purchase includes 30 days of direct setup support and 12 months of bug and security updates. If a material defect cannot be fixed, our refund policy provides a full refund.

Is this a subscription?

No. You pay once ($99 for your first mod, $75 for each additional mod) and can use and modify the source code in your product forever. Includes 12 months of bug and security updates, with zero recurring Helpful Mods fees.

How do I buy this mod?

Add this mod to your cart and pay through Stripe Checkout. Instant delivery: Download your complete source zip, SQL migrations, and agent setup prompts the second Stripe confirms checkout.

SHIP IT TODAY

Save weeks of plumbing on Website Chat.

Add Website Chat to your cart for $99 ($75 if it is an additional mod), or ask Mallow a technical question before you buy.