Two parts, both transparent: a monthly plan for the platform, and AI credits that map 1:1 to what the models actually cost. Every new workspace starts with 300 free credits — no credit card required.
Wondering how that compares? Every rival's list price, side by side — sourced, dated, and honest about what they do better.
The platform: inbox, bots, channels, order updates, broadcasts. The only line we set.
Every reply metered to the token at the provider's own price. 300 credits free to start.
Templates and order updates go to your own WhatsApp account's bill. We add nothing on top.
| What you get | starter — $29/mo | pro — $99/mo | scale — $299/mo |
|---|---|---|---|
| AI replies per month | 500 | 3,000 | 15,000 |
| WhatsApp broadcasts per month | 5,000 | 15,000 | Unlimited |
| WhatsApp order updates per month | 1,000 | 10,000 | Unlimited |
| AI credits included | 1,000 | 5,000 | runs on your own key |
| Replies from your team | Unlimited & free | Unlimited & free | Unlimited & free |
| Bots | 1 | 5 | 25 |
| Team members | 3 | 10 | 50 |
| Channels (website, WhatsApp, Messenger) | 1 | 3 | 10 |
| Standard models — Sonnet 5, GPT-5.6 Terra, Gemini Flash, DeepSeek, Qwen, Llama | ✓ | ✓ | ✓ |
| Flagship models — Opus 5, GPT-5.6 Sol, Grok 4.6, Gemini 3.1 Pro | — | ✓ | ✓ |
| Frontier models on your own key — Fable 5.1, GPT-6 Astra | — | — | ✓ |
| Knowledge base, website crawl, hosted help centre | ✓ | ✓ | ✓ |
| Live inbox, human handoff, contacts | ✓ | ✓ | ✓ |
| Email transcripts & customer requests | ✓ | ✓ | ✓ |
| Team roles, MCP for Claude, order updates API | ✓ | ✓ | ✓ |
| Store actions — order lookup, add to cart & checkout | — | ✓ | ✓ |
| Remove “Powered by Tendabot” | — | ✓ | ✓ |
| Bring your own API key (BYOK) | — | — | ✓ |
| Pack | Price | Works out at |
|---|---|---|
| +1,000 AI replies | $39 | $0.039 per reply |
| +5,000 AI replies | $149 | $0.030 per reply |
Your plan's included replies are always used first — a pack is only touched once they run out, so you never burn something you paid extra for.
| Top-up | Credits | Roughly |
|---|---|---|
| $25 | 2,300 credits | $23.00 of model usage |
| $50 | 4,600 credits | $46.00 of model usage |
| $100 | 9,200 credits | $92.00 of model usage |
Prices include payment processing. Unused top-ups roll over. 1 credit = $0.01 of real model cost — we don't mark up tokens. Need volume pricing or an invoice?
| Model | ≈ credits per reply | 1,000 replies | Runs on |
|---|---|---|---|
| Claude Haiku 4.5 | 0.2 | $2.00 | Every plan |
| Gemini 3.8 Flash | 0.15 | $1.50 | Every plan |
| DeepSeek V4 Flash | 0.06 | $0.60 | Every plan |
| GPT-5.6 Luna | 0.05 | $0.50 | Every plan |
| Claude Sonnet 5 | 0.4 | $4.00 | Every plan |
| GPT-5.6 Terra | 0.46 | $4.60 | Every plan |
A typical support answer: 500 tokens in, 300 out. Your credits go exactly this far — there is no markup to move the numbers. All 73 models →
Your plan includes a monthly number of AI replies. What each reply costs is measured in tokens and priced from the provider's own rate card with no markup — that's what credits track. Pick a small model and your credits stretch far; pick a flagship and they go faster. Your ledger shows every line. Replies your own team sends from the inbox are always free and never counted.
Your bot stops answering and the chat goes straight to your team instead, with a friendly note to the visitor — and you see it on your dashboard. Buy a message pack, top up credits, or add your own provider key and it resumes instantly. Your own replies from the inbox always send, cap or no cap, and you are never billed past what you bought.
No — Meta bills WhatsApp messages separately, per message, at its own rate for each country and category, and Tendabot adds nothing on top. From 1 October 2026 service replies (your answers inside the customer’s 24-hour window) are billed per message too, after a free 1,000 per phone number each month — roughly half a cent to two cents a message beyond that, depending on the country. Without a payment method on your WhatsApp Business account, Meta stops delivering at the 1,001st. Worked example in our WhatsApp pricing guide.
Starter runs every standard model — Sonnet 5, GPT-5.6 Terra, Gemini Flash, DeepSeek, Qwen, Llama and more. Pro adds the flagships such as Opus 5, GPT-5.6 Sol and Grok 4.6. Scale runs the frontier models, Fable 5.1 and GPT-6 Astra, on your own API key. A bot pointed at a model its plan no longer covers answers on the default model until you upgrade.
Budget picks like Llama 4 Scout, Qwen Flash or Gemini Flash-Lite run a typical reply for a few hundredths of a credit. See the full table on the models page.
Yes — paste your API key on the Billing page (verified live, stored encrypted, write-only). That provider then bills your account directly and uses zero Tendabot credits.
Every workspace starts with 300 free credits and full features — no card. That’s real usage, not a time-boxed demo.