--- title: "samambai pricing" url: https://samambai.com/pricing/ date: 2026-10-03 author: samambai --- Three ways to hire, in US dollars: a sprint at US$ 1,200 for 2 weeks, a project from US$ 3,000, and a partnership from US$ 900 a month. Meta, the model, n8n, and hosting are not inside those numbers. Prices at samambai are three, in US dollars, and they are on this page before the call: a sprint at US$ 1,200 for 2 weeks, a project from US$ 3,000, and a partnership from US$ 900 a month. The number covers our work. It does not cover Meta, the model, n8n, or hosting. Those accounts stay in your name. ## Three ways to hire | Way | Price | What is included | What is not included | | --- | --- | --- | --- | | Sprint | US$ 1,200 per 2-week cycle | One focus. Direct contact. Cancel whenever you want | A second focus in the same cycle. A multi-month roadmap. Project support after the cycle. A vendor invoice | | Project | From US$ 3,000 | Scope, timeline, and price locked. A delivery every sprint. Support after delivery | A change outside the locked scope. An open monthly roadmap. A vendor invoice. A promise of revenue or of a rank in AI answers | | Partnership | From US$ 900 a month | A continuous roadmap. Priority. A monthly check-in | A whole product built in month one at that price. An on-call promise with a clock on it, because that is not in the plan. A vendor invoice | Priority, on a partnership, means the queue sits ahead of someone who only buys an ad-hoc sprint. It does not mean a dedicated full-time person, and it does not mean a response time measured in minutes. If you need that agreement, it has to be written into the scope. It is not implied by US$ 900. Support after delivery exists on a project. The length is not a magic 30 or 90 days. It is written into the locked scope, so you can see what you are paying for. We do not invent a lifetime warranty. How these numbers sit next to SaaS seats and public vendor rates, with the source named, is in [how much AI automation costs](/blog/how-much-does-ai-automation-cost/). ## What fits in a 2-week sprint One focus fits, if access to the system already exists. "Transform the company" does not. These are the right size when the API or the export is already in hand: - One n8n flow: a new order in the store, into a spreadsheet and a WhatsApp notice for the person who operates it, with an alert if the flow fails. - One WhatsApp agent cut: one number on the official API, one rule (confirm a time that is already on the calendar), and a handoff to a person with the history. Several rules, several numbers, and templates for several countries push the work to a project. The service is [WhatsApp AI agents](/whatsapp-ai-agents/). - One cut of an app built in Lovable: authentication and one path that already works in the preview, moved to an environment where you can charge. The whole SaaS, with data migrated and exceptions covered, is a project. The page is [Lovable to production](/lovable-to-production/). - A GEO diagnosis, or the first technical pack on a small site. A full implementation often takes 2 to 4 weeks, so it can be one sprint or two. The service is [GEO](/geo/). If a second system shows up mid-sprint, it goes to the next cycle or into a project scope. It does not get squeezed in. That is how US$ 1,200 stays a true price. Most automations go live in 2 to 4 weeks. Full projects, in 4 to 8 weeks. Those are our delivery windows, not Meta's window for approving a message template. ## What changes the price The floor is public. What moves the number up from that floor is scope, and it can be named before the proposal: - How many systems are in, and whether they have an API or only a screen. - How many WhatsApp numbers, and how many languages. - How many rules and how many exceptions. "Answer the question" is one rule. "Answer the question, check a slot, reschedule, and call the front desk if the person describes a symptom" is four. - Whether the data is sensitive (health, identity documents, payment). The design gets narrower. It does not get "a bit more AI." - Whether human handoff needs a queue, hours, and an after-hours note, or whether a number that already answers is enough. - How many message templates, and for which countries. That changes your Meta invoice more than ours, and it still changes the time to write and test. What does not change the price: the buyer's city. São Paulo, New York, Madrid, and Tel Aviv are on the same table. Delivery is in English or in Portuguese. ## What you pay vendors directly These items do not pass through our invoice. You contract the vendor, on your card, in your company. **WhatsApp.** Since July 1, 2025, Meta charges per delivered message, by category and by the recipient's country [1]. The HTML pricing page, updated September 30, 2026 and read on October 2, says the October 1, 2026 changes are in effect: a service message, and a utility template sent inside the 24-hour window, are charged. Each business phone number gets 1,000 free service messages a month. Unused messages do not roll over. The charge starts at message 1,001. That allowance does not cover utility templates or marketing [1]. The Markdown export of the same document, read the same day, still describes free-form replies and in-window utility as free, and still treats the October 1 update as upcoming [5]. The page on non-template messages still titles itself as an upcoming update, states the same October 1 charges, and publishes 0.68 US cents as the current utility and authentication rate for Brazil. On that page, a service message uses the same rate and has no volume tier [2]. Three official surfaces do not say the same thing. The invoice in WhatsApp Manager is the one that counts. A serious budget counts the charged message and treats the 1,000 as the HTML page's allowance, not as a reply that stays free forever. An eligible account billed in Brazil creates new WhatsApp Business accounts in reais from July 1, 2026, and has until June 30, 2027 to move the others [5]. You pay the recipient's market, not the market of your headquarters. A New York company texting Brazil, Spain, or Israel pays those rows. **The model.** The OpenAI API pricing docs, read on October 2, 2026, list gpt-6-luna, one of the low-cost models on the table, at US$ 0.10 per million input tokens and US$ 0.50 per million output tokens, short context, standard rate [4]. The page shows no publication date. A short reply lands in a fraction of a cent. The arithmetic is in the [cost guide](/blog/how-much-does-ai-automation-cost/). Client data does not train a public model. The data stays in your account. **n8n.** On the annual plan published when we read the page on October 2, 2026: Starter at €20 a month (2,500 executions) and Pro at €50 a month (10,000 executions). Annual is labeled 17% below monthly. One execution is the whole workflow, not each step. Starter stops at 5 minutes. Pro, at 40. Hosted data sits in the European Union, in Frankfurt. The Community Edition runs on your server: no cloud fee of that kind, and the server bill is yours [3]. **Hosting.** A VPS, Supabase, Cloudflare, or whatever you already use. We do not resell that invoice, and we do not publish a server price here, because the price belongs to the provider you pick. Three currencies sit next to each other on purpose: our work in dollars, n8n in euros, Meta in dollars or in reais depending on the account. Cross-border card taxes and fees are not on this page, because they depend on your card and your country. Read the card statement. Do not add only the list price. ## The first conversation It lasts 30 minutes. No commitment, and no deposit to hold a date. We look at the system, the channel, and the rule. If a SaaS you can already buy will do the job, we say so. If the path is a sprint, a project, or a partnership, the proposal comes out with scope, timeline, and price. You already know the order of magnitude, because it is above. The process after that is the same one we always use: diagnosis, strategy, execution, evolution. A delivery every 2-week sprint. ## What we guarantee, and what we do not We guarantee what depends on us: one focus per sprint; something delivered every cycle; on a project, scope, timeline, and price locked, with post-delivery support written into that scope; on a sprint, cancellation whenever you want; human review over the agents we use to build, test, and audit. We do not guarantee what does not depend on us alone: revenue, the number of conversations resolved, a citation in ChatGPT or Gemini, your number's quality rating at Meta, approval of a template's wording on a given date, or the day a vendor releases a credential. Be wary of a contract that locks those items as if they were a software delivery. Data: least access, data in your accounts, LGPD and GDPR practice, and no training of a public model on your data. ## When not to hire Do not buy the sprint if the problem needs five systems at once. That is a project, and pretending it fits in US$ 1,200 wastes both sides. Do not buy the US$ 900 partnership expecting a complete app in month one. Do not hire us if you only want a tool license: the vendor sells that for less, and we are not a reseller. Do not hire us if the condition is a guaranteed return. Do not hire us if the work is a presentation with nothing in production. Hire us if there is a flow, an agent, or an app to put live, and you want the price before you tell the story. The services are on the [blog](/blog/) and on the pages for [WhatsApp](/whatsapp-ai-agents/), [n8n](/n8n-automation/), and [Lovable](/lovable-to-production/).