Loading page
About this tool
Estimate a support chatbot's monthly token bill from conversation volume, RAG context and turn length, then compare it to agent cost.
A support chatbot's monthly bill is conversation volume × cost per conversation, and cost per conversation is driven by what gets resent on every single turn: the system prompt, the retrieved knowledge-base context, and the transcript so far. For n turns with system prompt s, retrieval r, user message u and reply a, total prompt tokens are n(s + r) + u·n(n+1)/2 + a·n(n−1)/2, and output tokens are n·a. This estimator applies that model, adds prompt-cache discounts and retrieval cost, then compares the result against the agent handling cost the bot deflects.
Open Chatbot Monthly Cost Estimator on AltFTool — it loads instantly in your browser.
Enter the values you already know.
Fine-tune the options to match your scenario.
Read the result and use it in your planning or reporting.
Counts retrieval and system prompt on every turn, not once per conversation.
Reports net saving, ROI and the exact break-even containment rate.
Separate cached-input rate and a per-conversation retrieval cost line.
It is set by tokens, not by seats. A four-turn RAG conversation with an 800-token system prompt and 2,000 tokens of retrieved context runs roughly 13,000 prompt and 900 output tokens; at $1 per million input and $5 per million output that is under two cents a conversation, so 50,000 conversations cost a few hundred dollars.
Because it is injected fresh on every turn, not once. Four turns with 2,000 tokens of retrieved context means 8,000 prompt tokens of retrieval alone, which typically dwarfs the user's own words. Cutting chunk count or chunk size is usually the single biggest lever on the bill.
Divide the monthly bot cost by (conversations × fully loaded agent cost per contact). When a conversation costs under two cents to answer and a human contact costs several dollars, the break-even containment rate is often well under 1% — the business case usually rests on answer quality and escalation handling, not on token price.
Yes, and more than in most workloads, because the system prompt and tool definitions are identical across every conversation. Cached input is billed far below fresh input, so a high cache hit rate on the static prefix removes most of that portion of the bill; freshly retrieved chunks and the growing transcript still bill at full rate.
Add the Chatbot Monthly Cost Estimatorwidget to your blog or website — free, responsive, no signup. Just keep the “Widget by AltFTool” credit link visible.
<iframe src="https://www.altftool.com/embed/widget/chatbot-monthly-cost-estimator"
title="Chatbot Monthly Cost Estimator — free AltFTool widget"
width="100%" height="640" style="border:0;border-radius:12px;overflow:hidden"
loading="lazy" referrerpolicy="no-referrer-when-downgrade"></iframe>
<p style="font-size:12px;margin:4px 0 0">Widget by <a href="https://www.altftool.com/tools/all/chatbot-monthly-cost-estimator?utm_source=embed&utm_medium=widget">AltFTool — free online tools</a></p>