AI Chatbot Cost in 2026 — What It Costs to Add AI Chat to Your Site or App

Adding AI chat to an existing website or web app costs $3,000 to build plus $60/month to run, per our calculator engine — calibrated against 600+ project quotes and public rate benchmarks. This page covers the honest scope: bolting AI chat onto a product you already have, which is what almost every buyer searching this term actually needs.

Get Your AI Chatbot Cost Estimate

Quick Answer

Adding an AI chatbot to your website or app costs $3,000 to build plus $60/month to run. That is our engine's modeled cost for integrating AI chat (prompt design, API wiring, testing — about 15 hours) into an existing site or app. A simpler, non-AI live-chat widget costs $200 to build and $15/month instead. Calibrated against 600+ project quotes and public rate benchmarks across 6 markets.

Worked example: AI chatbot integration, US freelancer

AI chatbot integration (prompt design, API wiring, testing)15h
Total15h
× $135/hour (US freelancer, custom-feature rate band (median))
= $2,025

Hours are the engine baseline for this feature. Agency rates on the same 15 hours (median $232/hr, US) land closer to $3,480 — the $3,000 headline on this page sits between the two. Estimate your project →

Email me this AI chatbot cost breakdown ($3,000 + $60/mo)

No spam, ever. Unsubscribe anytime.

What actually makes up the $3,000

Our engine prices AI chat as a feature add-on, not a separate product. It adds about 15 hours of build work to whatever project you are already scoping — a new site, an existing web app, or an MVP — covering:

One-time build

$3,000

~15 hours: API integration, conversation-flow design, testing

Ongoing

$60/mo

LLM API usage + hosting the chat logic, at typical traffic

In our complexity engine, choosing an AI chatbot over basic live chat adds 10 points to a project's complexity score (of 100) instead of 1 — the integration and testing surface is real, not just a line item. A simpler, rule-based live-chat widget (no LLM) is also in the engine at $200 to build and $15/month, if AI chat is more than you need right now.

This scope is adding chat to a product you already have. Building an AI chatbot product from scratch — your own hosted model, a standalone SaaS around it — is a different, larger project we do not model on this page.

Why AI API costs are moving in 2026

Last reviewed: 2026-07-22

This is external market context, not our own benchmark data — a snapshot of published vendor pricing, cited by source. Five major model releases repriced the AI-API market inside the last three weeks:

ModelReleasedPrice (in/out per M tokens)Source
Gemini 3.6 Flash / 3.5 Flash-Lite / Flash Cyber2026-07-21$1.50 / $7.50 per M tokens (in/out)marktechpost.com, artificialanalysis.ai
Kimi K3 (open-weight, 2.8T)2026-07-16$3 / $15 per M tokens (in/out)the-decoder.com, simonwillison.net
GPT-5.6 family, incl. "Luna" tier2026-07-09$1 / $6 per M tokens (Luna, in/out)benchlm.ai
Meta Muse Spark 1.1 (Meta’s first paid API)2026-07-09$1.25 / $4.25 per M tokens (in/out)techtimes.com
Claude Fable 5 (added to Max plans)2026-07-01 / 2026-07-21$10 / $50 per M tokens (in/out)Hacker News

The genuine takeaway is not which model is cheapest this week — it is that sticker token price barely moves your total chatbot bill. The LLM API is only about 8–15% of total cost; the rest is the integration, UX, and testing work above, plus ongoing hosting. Even a 5× swing in per-token pricing (Kimi K3's jump from its predecessor is close to that) only shifts a small slice of the $60/month, not the $3,000 build.

Pricing news moves faster than the rest of this site — we are reviewing this table on a shorter cycle than our usual content and will update or retire it if it goes stale.

Frequently Asked Questions

How much does it cost to add an AI chatbot to my website or app?

Adding AI chat to an existing website or web app costs $3,000 as a one-time build plus $60/month to run, per our calculator engine. That covers integration, prompt and conversation-flow design, and testing — roughly 15 hours of development work. It assumes you already have a site or app to add the chatbot to, not a from-scratch product build.

How long does it take to add AI chat to an existing site or app?

Most integrations take 2–4 weeks: connecting to an LLM API, designing the conversation flow for your use case (support, lead qualification, product Q&A), and testing edge cases before launch. Timelines stretch if you need the chatbot trained on a large private knowledge base or wired into several internal systems.

Is the LLM API itself expensive?

No — the raw API line item is usually a small share of the total. Vendor pricing swings fast (see the July 2026 repricing wave below), but the LLM API is only about 8–15% of total chatbot cost. Most of the $3,000 + $60/month is integration, UX, testing, and ongoing hosting — not the tokens themselves.

What is the difference between a chatbot and a full AI agent?

A chatbot answers questions and follows a conversation script — the $3,000 scope on this page. An AI agent takes actions on a user’s behalf (booking, refunding, updating records), which needs tool access, permissions, and more testing, and costs more. See our dedicated AI agent cost breakdown for the full numbers ($8,775–$15,080, composited from real engine line items) if you need agent-style actions.

Read the AI agent cost breakdown →

Do I need to train my own AI model?

Almost never. Nearly every AI chatbot we scope integrates an existing model through an API (OpenAI, Anthropic, Google, or an open-weight model like Kimi) rather than training one from scratch. Training your own model costs orders of magnitude more and only makes sense at a scale most sites never reach.

What does the $60/month actually pay for?

The ongoing $60/month in our engine covers LLM API usage at typical traffic and the hosting/infrastructure for the chat logic itself. It scales with conversation volume — a low-traffic site may run under that, and a high-volume support widget can run well above it.

Can I add a simpler chatbot for less?

Yes. Our engine also prices a basic live-chat tier (rule-based or simple canned responses, no LLM) at $200 to build and $15/month to run — a fraction of the AI chatbot cost. It is a reasonable starting point if you mainly need quick-reply support and are not ready to commit to an AI-driven build.

Does adding AI chat make the rest of my project more complex?

Yes, measurably. In our complexity engine, the AI chatbot option adds 10 points to a project’s complexity score (on a 100-point scale) versus 1 point for basic live chat — the API integration, conversation-state handling, and testing surface add real project risk, not just a $3,000 line item.

Ready to price your AI chatbot?

Configure your real project and see where the $3,000 + $60/mo lands inside your full build.

Get Your AI Chatbot Estimate

Related Cost Calculators