Getting Started


## What is Ask Steve? 
Ask Steve is a Chrome extension that unlocks the power of LLMs like ChatGPT and Gemini in any web page. LLMs give you superpowers and Ask Steve adds them to every web page, so you can get your everyday tasks done better and faster. **And best of all, it's FREE!** Give it a try and see what you think. 

## Who is Steve?
* Steve is your AI intern who is very good with words
* He can write content for you, edit content for you, analyze content for you and much more
* If your job involves any of these things, Steve can help you do it faster
* And he can do it right in the web pages where you work, so you don't have to bounce back and forth between your work and an LLM like ChatGPT or Gemini

> And just like a human intern, you should always check his work! 👍

## Installing Ask Steve
<div style={{backgroundColor: 'violet', padding: '1rem'}}>
You can install Ask Steve in your Chrome Browser from the Chrome Webstore [here](https://chromewebstore.google.com/detail/ask-steve-do-your-job-bet/gldebcpkoojijledacjeboaehblhfbjg?hl=en&authuser=0). 
</div>

## Working with Steve
* Once you install the Ask Steve extension you'll be dropped onto the [onboarding page](https://quickstart.asksteve.to) that will teach you how to work with Steve. 
* If you want additional information, check out the docs on the left side of this page. 
* Be sure to pin Ask Steve to your Chrome toolbar for fast access!
<img src='/assets/images/posts/pin-extension.gif' alt="Pin Ask Steve"/>


Welcome to Ask Steve

Learn how Steve can help you get tasks done better & faster.

Working with Steve

Skills enable you to quickly ask Steve to do complex things

Skills

Select any text or page and get help from Steve

Context Menu

Talk about life, the universe, and the current page

Chat

Get help from Steve in any editable text field

Quick Action Box

Add permanent buttons to any page to make repetitive tasks FAST

Skill Buttons


## Why One-Time Requests?
[Skills](/docs/skills) are great for when you know ahead of time what you're going to need help with, but that's not the case a lot of the time. One-time requests enable you to ask for anything, at any time. And if you find that you're making the same one-time request over and over, you can always save it as a Skill!

Here are all the ways to make a one-time request with Ask Steve:
* Right-click to open the [Context-Menu](/docs/context-menu) on a page, on some selected text, or in an editable text field, and then select `Ask Steve to...` and then `ONE TIME REQUEST`
* Press the Ask Steve toolbar icon to open the sidepanel to access [Chat](/docs/chat), and then type in your question.
* Open the [Quick Action Box](/docs/quick-action-box), type in your request and then press `ENTER`



All the ways that you can make a one-time request to Steve

One-Time Requests


## Why Call an API or Webhook? 
* This gives you the ability to call your own APIs that can do pre- and post-processing. 
* As an example, you could configure your own internal API that Steve calls with an account number that he pulls from a page. Your API can use that account number to pull relevant data about that account from internal systems, directly make an LLM call with the new augmented data, and then pass the result back to Steve. 
* Or you could send the input to an LLM for processing, and then write it out to a Google Sheet or to Salesforce. 
* Or you could proxy all requests so that you can log them, or filter any sensitive information out. 
* You can do this with your own code running on a server somewhere, or with any visual workflow builder like [n8n](https://n8n.io/), [Gumloop](https://www.gumloop.com/), [Activepieces](https://www.activepieces.com/), [Make](https://www.make.com/), [Zapier](https://zapier.com/) etc. that supports Webhooks. 
* This enables you to do a lot more than just call an LLM! See the videos below for a few examples. 
<iframe width="560" height="315" src="https://www.youtube.com/embed/uUXwVwdVUgE?si=bdUYHPM0NkOIN2w3" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share; fullscreen" referrerpolicy="strict-origin-when-cross-origin" allowfullscreen></iframe>
<br />
<iframe width="560" height="315" src="https://www.youtube.com/embed/aNRfwC0bYzU?si=NoqQ32ryL_Xhndq_" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share; fullscreen" referrerpolicy="strict-origin-when-cross-origin" allowfullscreen></iframe>


## How to Call an API or Webhook
Ask Steve can interface with any API or Webhook on the back-end. 
* To point a Skill to an API/Webhook instead of an LLM, follow the instructions [here](/docs/skills#advanced-options) when setting up your Skill.
* API/Webhook Skills behave as follows: 
1. All requests are made via POST
2. In the Skill configuration, the default header is to set the Content-Type to "application/json", but you can add anything else to the header that you need, including an API key or Authentication key. 
3. The body of the POST request contains a property for every [Page Variable](/docs/skills#page-variables-in-skills) and [User Variable](/docs/skills#user-variables-in-skills). All other text in the "Instructions for Steve" area are ignored. 
4. A "_url" property is automatically added that has the URL of the current page. 
5. Your API/Webhook should return an object with a "message" property (string) if it was successful, and an "error" property (string) if it was not. This is what will be shown to the user. 
6. You can include simple HTML formatting in the return values. URLs will be hyperlinked automatically, and text can be hyperlinked by using [Markdown formatting](https://anvilproject.org/guides/content/creating-links). 

> **Example Instructions:** *This Skill captures [[leads]] and sends them to a Webhook*
>
> The POST body will have two properties, "leads" with the selected text or current page, and "_url" with the url of the current page. All other text is ignored and not sent to the API/Webhook. 

> **Example Instructions:** *This Skill captures a [[headline]], [[author]] and [[article]] then asks the user for \{\{what do you want to do with this\}\} and then sends them to a Webhook*
>
> The POST body will have five properties, "headline", "author", "article", "what do you want to do with this", and "_url". All other text is ignored and not sent to the API/Webhook. 


Calling APIs and Webhooks

Settings


## Accessing the Ask Steve Login page
You can access Settings in the following ways: 
* Right-click on the Ask Steve toolbar icon and select `Options`
* Click on the Ask Steve toolbar icon to open the sidepanel, then press the Gear icon in the upper right
* Following [this link](chrome-extension://gldebcpkoojijledacjeboaehblhfbjg/options.html)

## Sign Up / Login
* Here's where you can see your current plan and can create an account or login. You'll need to login in order to upgrade to Standard. 
* You can create an account with an email and password, in which case a confirmation email will be sent to you. 
* You can also login with your Google account, in which case no email confirmation is needed. 

## Setting Up Your First Model
* Once you login you can setup your first model. This will switch you from the QuickStart plan to the Standard plan.
* You can get your **FREE** Gemini API Key by following the instructions [here](https://quickstart.asksteve.to/getkey.html). **Free Gemini is not available in the EU, UK or Switzerland** ([details](/docs/where-it-works#countries)). If you are in one of those regions, you will need to use a different LLM provider or a [Local Model](/docs/local-models). 
* Alternately you can connect to one of the other models from OpenAI or Anthropic from the list. 
* Go to the [Models](/docs/models) page for more templates and to add/edit/delete models. 



Login to Ask Steve and set up your first model

Login


To access the Settings page, click on the `Settings` link in the left navigation of the Extension Settings.  Or [click here](chrome-extension://gldebcpkoojijledacjeboaehblhfbjg/settings.html). 

## Change the Ask Steve Sidepanel Theme
* **Light:** Use the light theme
* **Dark** Use the dark theme
* **System:** Use whatever theme your OS is set to use

Whenever you close and reopen the sidepanel, it will check for the latest theme setting and use it. 

## Show the Ask Steve Floating Action Button (FAB)
* By default Ask Steve shows a floating action button (FAB) on every page that gives you quick access to the Ask Steve sidepanel.
* Pressing the button opens the Ask Steve sidepanel.
* You can grab the drag handle on the FAB up and down the right edge of the page and Steve will remember where you placed it.
* If you'd like to turn off the FAB, uncheck the `Show FAB` checkbox and press `SAVE`


To access the Skill Library page, click on the `Skill Library` link in the left navigation of the Extension Settings. Or [click here](chrome-extension://gldebcpkoojijledacjeboaehblhfbjg/library.html).

## Installing a Skill
* Click the tile for any Skill that you're interested in. You can review the author's notes and the instructions for Steve.
* If you want to install the Skill, click on the `INSTALL` button.
* The Skill will be added to your Skills list. If you want to edit it, go to [Skills](/docs/skills) tab in the Ask Steve sidepanel. 

\
**NOTE:** Existing Skills are never overwritten. If there is an existing Skill with the same name, the one that you are adding will get a number after it, like New Skill (2). 


Find and install new Skills, then customize them and make them yours!

The Skill Library

Import, export and delete your Skills and Skill Buttons

Import, Export & Delete

<div style={{backgroundColor: 'violet', padding: '1rem'}}>
**You need to have your own paid account with the LLM service/model provider.**
</div>


To access the Models page, click on the `Models` link in the left navigation of the Extension Settings. Or [click here](chrome-extension://gldebcpkoojijledacjeboaehblhfbjg/models.html).

## What Does the Model Do? 
* The model is the LLM service that Steve uses to help you accomplish all your tasks. By default Ask Steve uses Google's Gemini 1.5 Flash model.
* All models have different characteristics and are better or worse at various tasks.
* Generally the better the model, the slower it and more expensive it is. 
* **When you use the free Gemini 1.5 Flash model, Google reserves the right to use your data for training purposes and to make their products better.** If your data is sensitive or you are concerned about privacy, you should create a paid account at a provider that doesn't use your data and connect to it. 
* Switching out Steve's model is like switching out his brain. 

## QuickStart Model
* The QuickStart plan uses Google's Gemini 1.5 Flash model.
* When you are on the QuickStart plan, your requests are sent to the Ask Steve server, which sends them along to Google, and then sends the results back. Your requests go through our server, but we never save anything. 

## Add a New Model
* To add a new model, press the `ADD NEW MODEL` button. A dialog will open with a number of model templates that you can select from, and a place for you to enter your API Key for that model if you have it already. Press `OK` and the interface will be filled in with the template details. 
* Edit any of the details that you need to
* To test the model, scroll down to the `Test Your Connection` area and press the `TEST` button. A sample query will be sent and the response shown below. 
* Once you are satisfied that everything is working correctly, press the `SAVE NEW MODEL` button. Steve will switch to using this new model, as indicated in the dropdown list at the top of the page. 

All requests are made via POST. The data fields for a model are as follows: 
* **Name:** So that you know what model this is. This is shown to users in error messages. 
* **URL:** The URL endpoint for this model
* **Context Window:** The size of the context window for this model.
* **Header:** The JSON header that is sent in the POST request. This should be formatted according to the service provider's documentation. This typically includes the API Key or other authentication for the service. 
* **Body:** The JSON body that is sent with the POST request. This should be formatted according to the service provider's documentation. 
* **Response Path:** In bracket notation, the path to the field in the response object that contains the actual response. This will be in the service provider's documentation. Note that one-shot responses and streaming responses usually have different paths, even from the same service provider. 
* **Error Path:** In bracket notation, the path to the field in the response object that contains any error messages. This will be in the service provider's documentation.
* **This is a Streaming API:** If you are calling a streaming API, this needs to be checked, as streaming responses are parsed differently than non-streaming responses. 

## Edit a Model
* Select the model that you want to edit from the `Steve is using:` dropdown. 
* Make your edits
* Press the `UPDATE` button
\
\
**NOTE:** This also changes the current model. Change it back when you're done editing via the `Steve is using:` dropdown. 

## Change the Current Model
* Select the model that you want to change to from the `Steve is using:` dropdown. 
* This is automatically saved, so now Steve will use this model.


## Delete a Model
* Select the model that you want to delete from the `Steve is using:` dropdown. 
* Press the `DELETE` button and confirm deletion
* Select whatever model you want Steve to use now

## Pro-Tips
### Call Any API
* If you want to call an API or Webhook from a Skill, see the details [here](/docs/webhooks)

### Change the Temperature
* The `temperature` setting of a model usually ranges from 0.0 to 1.0 or 2.0, where 0 tells the model to provide more factual answers, and the upper limit tells the model to provide more creative responses. The value can be set to a value somewhere in that range (e.g. 0.3)
* The default model and all our templates use a temperature of 0, as most use cases will benefit from more factual responses. 
* When adding a new model, you can set its temperature to something else, depending on how you're going to use it. 
* You can even add the same model multiple times, each with a different temperature, and then assign Skills that need more creativity to higher temperature models. 

### Connect to a Model Running On Your Own Machine
* See details on how to set this up [here](/docs/local-models)

Connect to other models where you have accounts

Connecting to Other Models

## What is a Local Model?
When you run a local model, the LLM runs on your own computer. Then Steve can use that model instead of using one in the cloud. The primary benefit of this is that **your requests never leave your machine, so all your data is kept private**. The secondary benefit is that it's FREE. We have tested Ask Steve with [LM Studio](https://lmstudio.ai/) and [Ollama](https://ollama.com/), but you should be able to use any system that can expose a model via a local web server.

### What are the downsides? 
* You have to download big model files (e.g. 3+GB)
* You need to have a fast machine
* The models your local machine can run generally aren't as good as the ones you can access in the cloud
* It can be slower, even though it's running locally

## How Do I Use LM Studio With Ask Steve?
1. Download and setup [LM Studio](https://lmstudio.ai/)
2. Open LM Studio. From the home page download a model. Try `Llama 3 - 8B Instruct`.
3. Once it's downloaded, go to the [Local Server](https://lmstudio.ai/docs/local-server) page in LM Studio and select the model you just downloaded from the dropdown at the top. If it doesn't start automatically, press the green Start Server button. This will start up a local server on port 1234 with the selected model.
4. Finally, go to the [Models](chrome-extension://gldebcpkoojijledacjeboaehblhfbjg/models.html) page in Ask Steve Settings, press `ADD NEW MODEL` and select `Local: LM Studio` from the menu. You won't need an API Key since it's running on your own machine. Press `TEST` to ensure it works, then `SAVE NEW MODEL` to save it. 
5. Congratulations! Steve is now using a model running completely on your own computer!


## How Do I Use Ollama With Ask Steve?
1. Download and setup [Ollama](https://ollama.com/)
2. Download a model. For example, this terminal command will pull Meta's Llama 3: `ollama pull llama3`
3. You will need to enable the Ask Steve extension to connect to Ollama. To do this you need to configure the Ollama server with the environment variable OLLAMA_ORIGINS, and it needs to be set to "chrome-extension://gldebcpkoojijledacjeboaehblhfbjg”.
4. Instructions for how to do so on various platforms are [here](https://github.com/ollama/ollama/blob/main/docs/faq.md#how-do-i-configure-ollama-server). So on a Mac you'd issue this terminal command: `launchctl setenv OLLAMA_ORIGINS "chrome-extension://gldebcpkoojijledacjeboaehblhfbjg"` to give Ask Steve access to Ollama. 
5. After setting OLLAMA_ORIGINS, you will need to restart the Ollama server. On a Mac you can do this by quitting and restarting Ollama from the Mac Taskbar. 
6. Finally, go to the [Models](chrome-extension://gldebcpkoojijledacjeboaehblhfbjg/models.html) page in Ask Steve Settings, press `ADD NEW MODEL` and select `Local: Ollama - Meta Llama 3` from the menu. You won't need an API Key since it's running on your own machine. Press `TEST` to ensure it works, then `SAVE NEW MODEL` to save it. Note that this assumes you downloaded the Llama 3 model. Change the "model" parameter in the body if you downloaded a different model. 
7. Congratulations! Steve is now using a model running completely on your own computer!




Connect to a model running on your local machine via Ollama or LM Studio

Connecting to a Local Model

<div style={{backgroundColor: 'violet', padding: '1rem'}}>
This is an Experimental feature. Please feel free to try it out, but YMMV. It seems to work best with GPT-4 class models (GPT-4o, Claude 3 Opus).
</div>


To access the memory page, click on the `Memory` link in the left navigation of the Extension Settings.  Or [click here](chrome-extension://gldebcpkoojijledacjeboaehblhfbjg/memories.html). 

## What is Memory?
* You can tell Steve facts or instructions that you want him to always keep in mind. 
* An example of a fact might be *My name is Inigo Montoya and the six-fingered man killed my father*
* An example of an instruction might be *Always respond in Spanish*
* An example of something more involved might be your whole resume.
* Every request you make to Steve includes all your memories. This way you can tell Steve something once and he'll "know" it every time he helps you out.  

> For example: If I save a memory with my name, job title and company, then in any Skill I can ask Steve to use them and he will know what they are. 

## Saving a Memory
* You can select any text on a page, open the [Context Menu](/docs/context-menu), select `Ask Steve to...` and then `REMEMBER THIS` at the bottom of the list. 
* In [Chat](/docs/chat) you can type `remember` followed by what you want Steve to remember (e.g. *remember that I don't like hot sauce*). Alternately if you have some text selected on the current page, you can can just type in `remember` and it will remember the selected text. 
* On the [Memory Page](chrome-extension://gldebcpkoojijledacjeboaehblhfbjg/memories.html) press the `ADD NEW MEMORY` button, type in your memory, and then save it. 

## Viewing, Editing and Deleting Memories
* You can do all these from the [Memory Page](chrome-extension://gldebcpkoojijledacjeboaehblhfbjg/memories.html)
* Your memories are shown in an alphabetized list. Select any one to view it, update it, or delete it. 

## Pro-Tips
* Try to keep memories short, as they take up some of the limited room in your LLM request
* Memory doesn't always work like you'd expect - some models are better than others at using them. You will need to experiment!




Have Steve remember facts and instructions

Memory (Experimental)

Other



## Websites are very... variable
Ask Steve will work on most websites, but given how powerful the web platform is and all the different ways that people can build sites and apps, it doesn't work everywhere. 
* It will most likely not work in web applications with complex layouts
* Sites that make excessive use of iframes and Shadow DOMs may cause problems

### Known Sites Steve Can't Help
* Google Docs
* Google Slides
* Figma
* XHTML sites, like the [SEC's document viewer](https://www.sec.gov/ix?doc=/Archives/edgar/data/1652044/000165204424000022/goog-20231231.htm). On sites like this, Steve does what you ask him, but then can't write his response to the page. You can always see his response by opening Chrome DevTools (Chrome > View Menu > Developer Tools) and then going to the Console tab and looking for the text where Steve is waving 👋. 

### Known Sites Steve Has... Issues
* LinkedIn: When composing a new Post, the Quick Action Box doesn't work.
* Salesforce: When sending an email, Salesforce doesn't "see" any of Steve's changes until you manually edit something in the email. So always be sure to edit Steve's work! 
* Facebook: Steve cannot write into the "Create post" dialog or Facebook comments. 
* Reddit: Steve cannot write into the comment field. New posts only work in the Markdown editor.


### Chrome Security Policies
* If you're using Ask Steve at work, your company may have set policies in Chrome that prohibit either the installation of certain extensions or the use of extensions on specific URLs. 
* In this case, you would need to ask a company administrator to add Ask Steve to the allowlist. 
* You can see what policies your browser is enforcing by opening `chrome://policy` and then looking for `Extension Settings`


## Countries
Ask Steve's Standard plan relies on the free Google Gemini API. Supported countries & territories are listed [here](https://ai.google.dev/gemini-api/docs/available-regions#available_regions). Any countries that have a [1] after their name **DO NOT support the Gemini FREE tier, so you will not be able to use the free Standard plan**. This includes the European Economic Area, Switzerland, and the United Kingdom. You will need to use another provider, pay for a Gemini account, or use a [Local Model](/docs/local-models) to connect to a model running on your own computer. 

## Browsers
Since Ask Steve is built for Chrome, it will at least partially work in most other Chromium-based browsers.
* **Chrome:** Works great. Install from the [Chrome Web Store](https://chromewebstore.google.com/detail/ask-steve-do-your-job-bet/gldebcpkoojijledacjeboaehblhfbjg?hl=en&authuser=0)
* **Edge:** In progress
* **Vivaldi:** Mostly works. The FAB to open/close the sidepanel does not. Install from the [Chrome Web Store](https://chromewebstore.google.com/detail/ask-steve-do-your-job-bet/gldebcpkoojijledacjeboaehblhfbjg?hl=en&authuser=0)
* **Brave:** Untested
* **Opera:** Untested
* **Arc:** Does not support the sidepanel so you cannot [Chat](chat) with a page or create or edit [Skills](skills) or [Skill Buttons](skill-buttons). You can still use the [Context Menu](context-menu) and the [Quick Action Box](quick-action-box). You can try to use a supported browser to create/edit your Skills and Skill Buttons, [export](export#export) them and then [import](export#import) them into Arc. Install from the [Chrome Web Store](https://chromewebstore.google.com/detail/ask-steve-do-your-job-bet/gldebcpkoojijledacjeboaehblhfbjg?hl=en&authuser=0)
* **Safari:** Not supported
* **Firefox:** Not supported

Ask Steve works on most websites, but not all of them

Where Does Ask Steve Work?


## Your Privacy
We take your privacy very seriously.
* We don't track what websites you visit
* We don't track what you're asking Steve for
* We don't track the Skills that you create
* We don't store your API Keys on our servers
* You can read our full privacy policy [here](/privacy).

## Our Philosophy
* Respect the user's privacy (see above)
* Try to keep Ask Steve servers out of the picture as much as possible
* Store as much data on the user's machine as possible
* Connect directly to the LLM provider whenever possible
* Be as transparent as possible about your data, how we use it, and where we send it (see below)
* [Here is the LinkedIn profile](https://www.linkedin.com/in/rajatrocks/) of the founder/creator of Ask Steve, you can decide if he looks trustworthy or not :)


## Extension Permissions
* When you install Ask Steve it will ask for permission to *"Read and change all your data on all websites"*. It needs this permission in order to be helpful on any web page. If you are uncomfortable with this then you should press `Cancel` and not install the extension. 
* If you want to limit what sites Ask Steve can work on, you can go into `chrome://extensions`, press the `Details` button on the Ask Steve tile, and then under `Site access` control what sites you want to give Steve access to.
* At any time you can disable Ask Steve via the toggle on the Manage Extensions page and enable it manually on sites you want to use it on.
* In an enterprise deployment, your administrator can configure certain sites to be allowed/disallowed via the `runtime_allowed_hosts` and `runtime_blocked_hosts` keys in [ExtensionSettings](https://support.google.com/chrome/a/answer/9867568?hl=en). You can use this to prevent Ask Steve from having access to sensitive internal pages. 

## Saved Data
### Your API Keys
* Ask Steve asks you to enter your LLM provider API keys so that it can connect to those providers.
> These keys are stored in the extension's [Local Storage](https://developer.chrome.com/docs/extensions/reference/api/storage) on your computer and are not sent to Ask Steve's servers.
* The API keys are only sent to the service provider in order to authenticate your requests. 
* The API Keys are not encrypted and are sent with every request to the service provider, so someone else with access to your computer could discover them. 

### Other Data
All other configuration data is stored in Local Storage on your computer. It is never stored anywhere else. This includes: 
* Skills
* Skill buttons
* Model configuration
* Chat history

> If you'd like to see what is being stored in Local Storage, do the following: 
> * Go into `chrome://extensions`
> * Enable `Developer mode` via the toggle in the upper right
> * Press the `service worker` link in the Ask Steve tile. This will open Chrome DevTools
> * On the Console tab, paste `await chrome.storage.local.get()` and then press `ENTER`
> * The contents of Local Storage will be printed out and you can inspect them

## Network Requests
### Your LLM Requests
* We send your request directly to the LLM service provider. 
* When you are on the QuickStart plan, your LLM requests are sent to to **quickstart.asksteve.to** which proxies them to Google's Gemini or another provider. The QuickStart proxy doesn't look at or store your request, it just passes it through. 

### Google Analytics
We log views of the extension pages (e.g. Settings, Models, Sidepanel, etc.), as well as the following information for every LLM request:
* Whether the call was successful or returned an error.
* The hostname for the service (e.g. `generativelanguage.googleapis.com` or `api.openai.com`)
* A transient session identifier
* This can be disabled for enterprise deployments
* Here is what that looks like for a call to the Google Gemini LLM:
```
{
  "name": "llm_success",
  "params": {
    "hostname": "generativelanguage.googleapis.com",
    "session_id": "1721321767437",
    "engagement_time_msec": 100
  }
}
```

### Login and Authentication
* Ask Steve uses Firebase (from Google) for login and authentication. Requests are made to Google's servers whenever you login or create a new account. 
* We don't ask for or store any PII beyond what is provided at login. 
* This can be disabled for enterprise deployments.

### The Skill Library
* Ask Steve uses Firebase for its backend database. Requests are made to Firebase to retrieve the list of Library Skills and to (in the future) upload a Skill to the Library.
* This can be disabled or pointed to an internal database for enterprise deployments. 

### Configuration Files
Ask Steve loads the following configuration files from **quickstart.asksteve.to**
* **models.json** - contains all the model templates
* **skills.json** - contains the out-of-the-box Skills
* **buttons.json** - contains the out-of-the-box Skill Buttons
* This can be disabled for enterprise deployments. 


> If you'd like to monitor all the network requests that Ask Steve is making, do the following: 
> * Go into `chrome://extensions`
> * Enable `Developer mode` via the toggle in the upper right
> * Press the `service worker` link in the Ask Steve tile. This will open Chrome DevTools
> * Switch to the Network tab. From here you can see all the network requests that Ask Steve makes

## Other
The LLM service may use your data for training or to make their products better. If you are on the QuickStart plan or are using the Google Gemini free tier, [Google does reserve the right to use your data](https://ai.google.dev/terms). If this concerns you, you should connect to a paid service or internal corporate service that explicitly states that they will not use your data. You can also conenct to a [Local Model](/docs/local-models) running on your own machine. 






What permissions does Ask Steve ask for and how does it use your data?

Permissions & Your Data


## QuickStart (Free!)
* When you first install Ask Steve you are put on our QuickStart plan. This gives you 100 requests that you can use to try out working with Steve and see if you find him helpful. 
* You can use all the functionality of the Standard plan, there is just a limit on the number of requests. 
* Responses from Steve come all at once, whereas on the Standard plan responses are streamed down as soon as Steve starts writing the answer, so you can work even faster. 
* Community support

## Standard (Free!)
* Ask Steve asks you to enter your own free Google Gemini API Key to enable the Standard plan. [See how to get your own key here](https://quickstart.asksteve.to/getkey.html). 
* You can alternately connect to any other provider where you have an API Key, or a [Local Model](/docs/local-models) running on your own machine.
* You continue to have access to all the functionality, with the only limits being what service provider imposes. The free Google Gemini API Key uses Gemini 1.5 Flash and has [the following limits](https://ai.google.dev/pricing).
* **Free Gemini is not available in the EU, UK or Switzerland** ([details](/docs/where-it-works#countries)). If you are in one of those regions, you will need to use a different LLM provider or a [Local Model](/docs/local-models). 
* Steve's responses are streamed, so you can work even faster.  
* Community support

## Premium (Coming Soon)
* Central team billing
* Central team user management
* Central team Skill Library
* Backup to server
* Custom Skill packs
* Email support


## Enterprise (Coming Soon)
* Integration services
* On-premise deployment
* Volume-based discounts
* Email support


Plans


> Want to revisit the onboarding pages? [Follow this link!](https://quickstart.asksteve.to)

Want to run through the onboarding again?

Onboarding Pages


> We made Ask Steve **completely free** because we wanted as many people to benefit as possible. 

We would appreciate any support that you can give us so that we can continue to make Ask Steve even better! Here are some ways that you can help us: 
* Writing a good review on the Chrome Web Store [here](https://chromewebstore.google.com/detail/ask-steve-do-your-job-bet/gldebcpkoojijledacjeboaehblhfbjg?hl=en&authuser=0)
* Sending us a testimonial about how you're using Ask Steve [here](https://forms.gle/KAEAeCLsGeLAGqKk8)
* Spreading the word and telling your friends and coworkers about Ask Steve
* Sharing tips, tricks and Skills in our [User Community](https://groups.google.com/g/ask-steve)
\
\
**Thank you!** 🙏🙏🙏




Support Us


## v1.1.3 - TBD
* Right click on the FAB/TAB (the purple tab with the lightning bolt icon) to temporarily remove it from a page, in case you need to print or take a screenshot. 
* Removed scrollbars on Windows on the Import/Export page. 


## v1.1.2 - August 27, 2024
* The Ask Steve lightning bolt tab/FAB no longer appears when you print pages. 
* Fixed issues with the format of text generated by the [Quick Action Box](/docs/quick-action-box) in Outlook 365. 


## v1.1.1 - August 16, 2024
* Many minor bug fixes
* Added better error messages if you try to use Ask Steve with Google Docs, Slides, Sheets or Figma, since it can't read content from any of those. 


## v1.1.0 - July 30, 2024
* Fixed bugs.  
* Changed how calls to APIs/Webhooks are made to make it easier - see the updated [documentation](/docs/webhooks). 
* As a result, removed the "Show Model Override" setting on the Settings page and the "This is Not an LLM" setting on the Models page since they're no longer needed. 
* Removed omnibox support


## v1.0.9 - July 24, 2024
* Automatically hide the FAB when a browser window is in fullscreen mode (so it doesn’t get in the way of your YouTube videos)
* Chat and the Floating panel now hyperlink all URLs
* Added a new “Show Model Override” setting on the Settings page that adds the ability to assign a specific model to a specific Skill. This is useful for when certain models are better at certain tasks, but also for when you want to call non-model APIs. See the video in the next bullet for an example.
* Added a “This is Not an LLM” setting on the Models page that indicates that you’re calling a non-LLM API. All the usual instruction text and Memories that Steve would send with your request are removed. See [this video](https://www.youtube.com/watch?v=uUXwVwdVUgE) for an example. 



## v1.0.8 - July 18, 2024
* OpenAI [made a change](https://community.openai.com/t/streaming-events-returned-bunched-up/870324) to how they send down streaming responses that broke our parsing. This has been fixed. 
* The sidepanel code has been changed so that there is one universal sidepanel for all the tabs in a window vs. a unique sidepanel per tab.
> * This change was made because unique sidepanels per tab is not supported in Edge and Vivaldi, and we weren’t really doing anything with that capability anyway. 
> * So now when you open the sidepanel, it is open on all the tabs in a window, and when you close it, it’s closed for all tabs. 
> * As a result, Chat and Skill configuration should now work properly in Edge and (mostly) in Vivaldi.
> * The lightning bolt tab that enables you to quickly open the sidepanel (aka the FAB) still doesn’t work properly in Vivaldi - you can disable it in the Settings.
> * If you encounter other issues, let me know here! 
* You can now open the Skill Picker in Chat with either  “k “ or just SPACE in the chat field. 
* When you go to the Models page, it is now more clear that you can’t enter model details until you login. Once you login, you can enter model details immediately without having to press “Add New Model”. 
* Added [Cohere Command R+](https://cohere.com/command) and Open AI GPT4o Mini as model templates. Removed OpenAI GPT 3.5
* Chat can use just the selected text on a page, but previously could not read selected text in iframes served from different domains as the main window. This has been fixed.
* For Skill Buttons - the output option “Popup” has been renamed “Floating panel” since some users thought it meant browser popup.


Notes for each release in the Chrome Web Store

Release Notes

"use strict";
const {Fragment: _Fragment, jsx: _jsx, jsxs: _jsxs} = arguments[0];
function _createMdxContent(props) {
  const _components = {
    a: "a",
    code: "code",
    h2: "h2",
    h3: "h3",
    li: "li",
    ol: "ol",
    p: "p",
    span: "span",
    strong: "strong",
    ul: "ul",
    ...props.components
  };
  return _jsxs(_Fragment, {
    children: [_jsxs(_components.h2, {
      id: "what-is-a-local-model",
      children: [_jsx(_components.a, {
        "aria-hidden": "true",
        tabIndex: "-1",
        href: "#what-is-a-local-model",
        children: _jsx(_components.span, {
          className: "icon icon-link"
        })
      }), "What is a Local Model?"]
    }), "\n", _jsxs(_components.p, {
      children: ["When you run a local model, the LLM runs on your own computer. Then Steve can use that model instead of using one in the cloud. The primary benefit of this is that ", _jsx(_components.strong, {
        children: "your requests never leave your machine, so all your data is kept private"
      }), ". The secondary benefit is that it's FREE. We have tested Ask Steve with ", _jsx(_components.a, {
        href: "https://lmstudio.ai/",
        children: "LM Studio"
      }), " and ", _jsx(_components.a, {
        href: "https://ollama.com/",
        children: "Ollama"
      }), ", but you should be able to use any system that can expose a model via a local web server."]
    }), "\n", _jsxs(_components.h3, {
      id: "what-are-the-downsides",
      children: [_jsx(_components.a, {
        "aria-hidden": "true",
        tabIndex: "-1",
        href: "#what-are-the-downsides",
        children: _jsx(_components.span, {
          className: "icon icon-link"
        })
      }), "What are the downsides?"]
    }), "\n", _jsxs(_components.ul, {
      children: ["\n", _jsx(_components.li, {
        children: "You have to download big model files (e.g. 3+GB)"
      }), "\n", _jsx(_components.li, {
        children: "You need to have a fast machine"
      }), "\n", _jsx(_components.li, {
        children: "The models your local machine can run generally aren't as good as the ones you can access in the cloud"
      }), "\n", _jsx(_components.li, {
        children: "It can be slower, even though it's running locally"
      }), "\n"]
    }), "\n", _jsxs(_components.h2, {
      id: "how-do-i-use-lm-studio-with-ask-steve",
      children: [_jsx(_components.a, {
        "aria-hidden": "true",
        tabIndex: "-1",
        href: "#how-do-i-use-lm-studio-with-ask-steve",
        children: _jsx(_components.span, {
          className: "icon icon-link"
        })
      }), "How Do I Use LM Studio With Ask Steve?"]
    }), "\n", _jsxs(_components.ol, {
      children: ["\n", _jsxs(_components.li, {
        children: ["Download and setup ", _jsx(_components.a, {
          href: "https://lmstudio.ai/",
          children: "LM Studio"
        })]
      }), "\n", _jsxs(_components.li, {
        children: ["Open LM Studio. From the home page download a model. Try ", _jsx(_components.code, {
          children: "Llama 3 - 8B Instruct"
        }), "."]
      }), "\n", _jsxs(_components.li, {
        children: ["Once it's downloaded, go to the ", _jsx(_components.a, {
          href: "https://lmstudio.ai/docs/local-server",
          children: "Local Server"
        }), " page in LM Studio and select the model you just downloaded from the dropdown at the top. If it doesn't start automatically, press the green Start Server button. This will start up a local server on port 1234 with the selected model."]
      }), "\n", _jsxs(_components.li, {
        children: ["Finally, go to the ", _jsx(_components.a, {
          href: "chrome-extension://gldebcpkoojijledacjeboaehblhfbjg/models.html",
          children: "Models"
        }), " page in Ask Steve Settings, press ", _jsx(_components.code, {
          children: "ADD NEW MODEL"
        }), " and select ", _jsx(_components.code, {
          children: "Local: LM Studio"
        }), " from the menu. You won't need an API Key since it's running on your own machine. Press ", _jsx(_components.code, {
          children: "TEST"
        }), " to ensure it works, then ", _jsx(_components.code, {
          children: "SAVE NEW MODEL"
        }), " to save it."]
      }), "\n", _jsx(_components.li, {
        children: "Congratulations! Steve is now using a model running completely on your own computer!"
      }), "\n"]
    }), "\n", _jsxs(_components.h2, {
      id: "how-do-i-use-ollama-with-ask-steve",
      children: [_jsx(_components.a, {
        "aria-hidden": "true",
        tabIndex: "-1",
        href: "#how-do-i-use-ollama-with-ask-steve",
        children: _jsx(_components.span, {
          className: "icon icon-link"
        })
      }), "How Do I Use Ollama With Ask Steve?"]
    }), "\n", _jsxs(_components.ol, {
      children: ["\n", _jsxs(_components.li, {
        children: ["Download and setup ", _jsx(_components.a, {
          href: "https://ollama.com/",
          children: "Ollama"
        })]
      }), "\n", _jsxs(_components.li, {
        children: ["Download a model. For example, this terminal command will pull Meta's Llama 3: ", _jsx(_components.code, {
          children: "ollama pull llama3"
        })]
      }), "\n", _jsx(_components.li, {
        children: "You will need to enable the Ask Steve extension to connect to Ollama. To do this you need to configure the Ollama server with the environment variable OLLAMA_ORIGINS, and it needs to be set to \"chrome-extension://gldebcpkoojijledacjeboaehblhfbjg”."
      }), "\n", _jsxs(_components.li, {
        children: ["Instructions for how to do so on various platforms are ", _jsx(_components.a, {
          href: "https://github.com/ollama/ollama/blob/main/docs/faq.md#how-do-i-configure-ollama-server",
          children: "here"
        }), ". So on a Mac you'd issue this terminal command: ", _jsx(_components.code, {
          children: "launchctl setenv OLLAMA_ORIGINS \"chrome-extension://gldebcpkoojijledacjeboaehblhfbjg\""
        }), " to give Ask Steve access to Ollama."]
      }), "\n", _jsx(_components.li, {
        children: "After setting OLLAMA_ORIGINS, you will need to restart the Ollama server. On a Mac you can do this by quitting and restarting Ollama from the Mac Taskbar."
      }), "\n", _jsxs(_components.li, {
        children: ["Finally, go to the ", _jsx(_components.a, {
          href: "chrome-extension://gldebcpkoojijledacjeboaehblhfbjg/models.html",
          children: "Models"
        }), " page in Ask Steve Settings, press ", _jsx(_components.code, {
          children: "ADD NEW MODEL"
        }), " and select ", _jsx(_components.code, {
          children: "Local: Ollama - Meta Llama 3"
        }), " from the menu. You won't need an API Key since it's running on your own machine. Press ", _jsx(_components.code, {
          children: "TEST"
        }), " to ensure it works, then ", _jsx(_components.code, {
          children: "SAVE NEW MODEL"
        }), " to save it. Note that this assumes you downloaded the Llama 3 model. Change the \"model\" parameter in the body if you downloaded a different model."]
      }), "\n", _jsx(_components.li, {
        children: "Congratulations! Steve is now using a model running completely on your own computer!"
      }), "\n"]
    })]
  });
}
function MDXContent(props = {}) {
  const {wrapper: MDXLayout} = props.components || ({});
  return MDXLayout ? _jsx(MDXLayout, {
    ...props,
    children: _jsx(_createMdxContent, {
      ...props
    })
  }) : _createMdxContent(props);
}
return {
  default: MDXContent
};


Getting Started

Working with Steve

Settings

Other

Connecting to a Local Model

Connect to a model running on your local machine via Ollama or LM Studio

What is a Local Model?

What are the downsides?

How Do I Use LM Studio With Ask Steve?

How Do I Use Ollama With Ask Steve?

Table of Contents