- Home
- Skills
- Agents & Orchestration
- Cerebro: AI Assistant for Amazon Echo in Argentine Spanish
Works with the AI Agents you already use
Cerebro: AI Assistant for Amazon Echo in Argentine Spanish
Alexa on its own answers what Amazon decided it should answer.
$10
Cerebro: AI Assistant for Amazon Echo in Argentine Spanish
Example session with this skill installed
Install Cerebro on my Echo. Just the skill for now, no PC agent. I live in Córdoba, Argentina.
- Read your context and instructions
- Compiled the cerebro
Summary
I prepared everything for the skill-only setup with Córdoba as your city. Three things are yours to do, because they happen inside your own accounts: creating the Gemini key, creating the skill in the Alexa developer console, and pasting the files below. It takes about 20 minutes and needs no credit card.
1. Get your Gemini key (2 minutes)
Go to aistudio.google.com/apikey, create a key and keep it to yourself. Do not paste it in this chat: you will type it directly in the console in step 4.
2. Create the skill
In developer.amazon.com/alexa/console/ask, with the same Amazon account your Echo uses, choose
Create Skill and set:
| Field | Value |
|---|---|
| Skill name | Cerebro |
| Primary locale | Spanish (US) |
| Type of experience | Other |
| Model | Custom |
| Hosting services | Alexa-hosted (Python) |
| Template | Start from scratch |
3. Interaction model and permissions
Interaction Model > JSON Editor: delete what is there and paste the whole content of cerebro/skill-package/interactionModels/custom/es-US.json. Then
Save Model and
Build Model.
2. Permissions: turn on Reminders and Timers.
3. Interfaces: turn on Audio Player, or the radio will not play.
The invocation name stays as "cerebro", so you will say "Alexa, abrí cerebro".
4. Code
In the Code tab:
- Replace
lambda_function.pywithcerebro/lambda/lambda_function.py. - Replace
requirements.txtwithcerebro/lambda/requirements.txt. - Create
lambda/config.pyand paste the content ofcerebro/lambda/config.example.py, changing only these lines:
GEMINI_API_KEY = "" # type your key here, in the console
CITY = "Córdoba"
LATITUDE = -31.4201
LONGITUDE = -64.1888
UTC_OFFSET_HOURS = -3
TIMEZONE_ID = "America/Argentina/Cordoba"
CEREBRO_LINK = "" # stays empty: no PC agent for now
- Save, then Deploy. The first deploy installs the dependencies and takes two or three minutes.
5. Check that it works
In the Test tab, switch from Off to Development and type:
| You type | Expected |
|---|---|
| abrí cerebro | "Hola. ¿En qué te ayudo?" |
| ¿a cuánto está el dólar? | Today's rates, read aloud |
| poné dos minutos | A two-minute timer |
| ¿qué tiempo hace? | The weather in Córdoba |
It then works on your Echo right away, without publishing anything: "Alexa, abrí cerebro".
If it does not answer
- Alexa says it cannot help, or opens something else: set the Echo's language to Español (Estados Unidos) in the Alexa app, and check the console shows a successful build.
- The first reminder asks for permission: accept the card that arrives in the Alexa app. It happens once.
Next steps
- Tell me what each test phrase answered and I will check anything that is off.
- When you want notes, lights, music and the control panel, ask me to add the PC agent. It needs Windows and Python 3.11 or newer, and takes about ten more minutes.
- Everything you can say is in
cerebro/docs/USO.md.
Connects securely to your tools. The creator never sees your data.
About this skill
Alexa on its own answers what Amazon decided it should answer. Cerebro connects your Echo to a language model and gives it three things it lacks: memory of what you tell it, actions, and hands on your computer. It is built for Argentina: it understands voseo, knows what the blue dollar and the country risk are, and knows the public holidays and the local football teams.
This skill has your coding agent install it for you, check that it works and adapt it to your home.
What you can say once it is installed
- "mi vieja se llama Marta" ... "¿cómo se llama mi vieja?"
- "recordame en veinte minutos sacar la pizza", "despertame a las siete", "todos los lunes a las ocho, la basura"
- "¿a cuánto está el dólar?", "¿cuándo es el próximo feriado?", "¿llueve el finde?", "¿cómo salió Boca?"
- "poné Mitre", "buenos días" (the day's summary)
- With the PC agent: "anotá que tengo que llamar a Marta", "prendé la luz del living", "poné música de Soda Stereo", "mandale un WhatsApp a Juan que llego tarde", "¿tengo mails?"
What your agent does
- Asks which setup you want: the skill alone, the skill with the PC agent, or two computers.
- Prepares every file and tells you exactly what to paste in the Alexa developer console. The console steps are yours; the agent cannot click there for you.
- Installs the PC agent on Windows, generates the signed link between the skill and your PC, and checks the log.
- Verifies each stage with real phrases, and reads the logs when something fails.
- Changes the assistant on request: new commands, new actions, a different invocation name. The package's tests tell it whether it broke anything.
What is in the package
- The Alexa skill (Python, Alexa-hosted) and its interaction model for es-US.
- The PC agent for Windows, a local web control panel, a desktop panel for a second computer and an on-screen orb.
- Guides in Spanish for installing and using it.
- 133 automated tests that run with no network, no AWS account and no Windows.
It costs nothing to run
The skill is hosted by Amazon, the model is Gemini's free tier and the live data comes from public key-free APIs. You need no credit card at any step.
Open source, and what this listing adds
Cerebro is open source under the MIT licence, and its source is free on GitHub at github.com/alfregarrone/cerebro. If you prefer to install it by hand, use the repository. This listing is the packaged release (built from a fixed commit, tests passing, checked to contain no keys or personal data) together with the workflow that lets your agent do the installation and the changes for you. Buying it also supports the project.
Safety
- Your keys stay on your machines. The package ships with empty templates, and the agent is told never to print or commit a key.
- The skill and the PC talk through outbound connections only, and every message is signed with HMAC-SHA256.
- The PC agent runs only a closed list of actions. Shutting down, sending a message and running a planned task ask you first.
- The control panel listens on your computer only, unless you choose to open it to your home network. It needs a token and must never be exposed to the internet.
Requirements
- An Amazon account and an Echo, or the Alexa app.
- A Google account for the free Gemini key.
- For the PC agent, the panel and the orb: Windows and Python 3.11 or newer.
Not covered
Publishing the skill to the Alexa store, and a PC agent for macOS or Linux. The guides inside the package are in Spanish.
PC agent requires Windows and Python 3.11+. Alexa skill restricted to es-US locale. Gemini free tier subject to daily rate limits. Alexa-hosted Lambda limited to Python 3.8.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 2 days ago
Requires
Permissions
Allowed Hosts
The agent reads the package, prepares configuration files and, for the optional PC agent, runs the installer and reads its log in a terminal. Once installed, Cerebro itself uses the network: the skill calls Gemini with your own key and public key-free data APIs (weather, dollar rates, holidays, news, football), and the skill and the PC exchange HMAC-signed messages through ntfy.sh using outbound connections only. Optional features you turn on yourself use Telegram, Gmail (read-only, with an app password), a Google Calendar iCal link, desktop WhatsApp and WiZ or TP-Link lights. Your keys stay in files on your own machines; the package ships with empty templates.